Introduction

Centres specialised in the treatment of people with eating disorders (EDs) have observed that appropriate treatment often occurs late, many months, or even years, after the appearance of the first symptoms [1]. Further to this, the longer the duration of anorexia nervosa (AN) with no treatment, the less efficacious is the treatment [2]. Indeed, it has been demonstrated that when AN is treated more than three years after the start of symptoms, the effects of the treatment are reduced [3]. A patient aged 19 or under at the time the eating disorder is detected is four times more likely to recover from AN, and eight times more likely to recover from bulimia nervosa (BN), compared to individuals detected after 20 years, probably because of the shorter duration of the evolution [4].

Early detection could help to reduce the duration of an untreated eating disorder (DUED), i.e. the time between the appearance of the symptoms and the start of appropriate treatment, and could consequently reduce associated complications. This time lapse is estimated at over two years for AN (29.9 months on average), four years for BN (53.0 months on average) [5], and more than five years for binge-eating disorder (BED) (67.4 months on average)[1]. This time lapse increases with age at first treatment [5]. In addition, a study with 22 years' follow-up showed that people with long-standing ED symptoms were significantly more likely to have had a longer DUED compared to those in remission [6].

Furthermore, the longer the duration of the illness, the higher the mortality rate [7]. The risk of a suicide attempt increases with the duration of the illness, especially because of social isolation, which worsens over time [8]. A decrease in mortality among female patients hospitalised for AN in the 80 s was observed in Sweden. The mortality rate was reduced more than threefold in a decade; this improvement in prognosis could be linked to a better detection and to an earlier start of treatment, following the development of specialised teams. Furthermore, there is more and more information on this pathology, and hospitalisations thus tend to occur earlier [9].

From a neurobiological point of view, it can be hypothesised that a longer DUED increases the duration of malnutrition. Secondary nutritional factors have an impact on comorbidities, which can, in turn, maintain EDs and thus aggravate the risk of neuro-progression and lead to a more serious prognosis [10]. Associations have been found between the duration of AN and cognitive functions: the longer the duration of AN, the greater risk of grey matter atrophy in the cerebellum and the midbrain [11], and the greater the risk of a decrease in cognitive performance [12]. Correlations have also been evidenced between the duration of AN and emotional and social functioning: the longer the duration, the fewer the positive facial expressions observed [13], the more severe are the anomalies in social functioning [14], and the greater the exhaustion of family and friendly resources, progressively leading to a social network deficit [15]. Social withdrawal can also be found in BN and BED, connected to a feeling of shame caused by hyperphagia and purging behaviours and the negative representations of obesity [16, 17].

From an economic point of view, beyond the human cost, EDs lead to various types of cost [18]. Indeed, there are costs relating to health system expenditure, and costs resulting from losses in productivity for people with an ED. Furthermore, informal care provided by family members and others also has a value. Finally, we can add costs of government and non-government programs, assistance, transport and accommodation costs associated with receiving treatment, and funeral costs.

In 2012, the total costs generated by EDs were estimated at 34 billion euros in Australia [18]. In comparison, in the same country, anxiety and depression accounted for 26.8 billion euros, obesity for 34.4 billion euros and consequences of cigarette smoking for 183 billion euros.

An American report published in 2020 estimates at 64.7 billion dollars the yearly economic cost of EDs [19]. Similar results are found in the United Kingdom [20]. Thus, earlier care would enable greater efficacy in the treatment of EDs, which could reduce the number and the duration of hospitalisations and improve productivity at work. Alongside, at present only 30 to 50% of people with an ED seek treatment [21]; thus, early detection could lead to a short term increase in health costs but a decrease in health and social expenditure in the long-term [18].

EDs mainly begin during the adolescent years [22]; the population at risk is found in schools and sporting populations [23] and more widely in all fields of healthcare [24]

To summarise, earlier care enables an aggravation of the somatic, psychiatric and social aspects of EDs to be avoided. In the evolution of EDs, the early phase is defined as less than 3 years of evolution [2, 10]. The early detection approach is part of secondary prevention: secondary prevention in mental disorders targets individuals meeting diagnostic criteria in the early stages of the illness; the aims of prevention are early detection and intervention for these patients [25]. This review focuses on the first phase of secondary prevention (early detection) and does not cover the early treatment phase.

International guidelines [24, 26,27,28] have emphasised the importance of early detection of EDs. This review will focus on both EDs and disordered eating, as it is well known that in adolescence, disordered eating has a similar negative impact to that of established diagnosis of EDs among adults health [29,30,31]. The means to achieve early detection are not detailed in the guidelines. Although a recent review about primary prevention exists [32], there has so far not been any review collecting data on articles concerning early detection. The aim of this review comprising an inventory of all the research focused on early detection of EDs, is to detail the types of study that exist, on which populations, their limitations, and their results, to highlight which types of study still need to be developed. The authors chose to perform a scoping review rather than a more focused systematic review. In fact, this subject has generated very few methodologically sound studies, and it did not seem appropriate to narrow the research field to a single question given that it is the first review on the subject.

Methodology

The literature review was carried out following the PRISMA-ScR method [33]. There is no review protocol already registered.

Eligibility criteria

The review concerned secondary prevention of EDs in terms of early detection.

The inclusion criteria were: articles in English or French published up to January 2021, concerning secondary prevention in terms of detection/screening, focusing on EDs, and published in peer-reviewed journals (see below for details). We included all types of study.

The exclusion criteria were: studies on primary prevention, on secondary prevention focusing on early intervention (and not detection) and on tertiary prevention.

Information sources and searches

This scoping review was conducted by two researchers, N.K. and N.G., working blind, on the PUBMED and PsycINFO databases including all dates up to January 2021. The search algorithm used in PUBMED was ("eating disorder*"[tiab] OR "anorexia nervosa"[Title/Abstract] OR "bulimia nervosa"[Title/Abstract] OR "binge eating disorder*"[Title/Abstract] OR "disordered eating"[Title/Abstract] OR "appetite disorder*"[tiab]) AND ("Secondary Prevention"[Mesh] OR "detect*"[Title/Abstract] OR "identification"[Title/Abstract] OR "recogni*"[Title/Abstract] OR "secondary prevention"[Title/Abstract] OR "diagnos*"[tiab] OR "mental health literacy"[tiab]) AND (english[Filter] OR french[Filter]). The same terms were used on PyscINFO. The screening and selection of articles is shown in Fig. 1. The decision to include or exclude studies was initially made on the basis of the article title, then its abstract, and finally the full text. The researchers then discussed their choices and came to an agreement. Duplicates were excluded. A manual search was also conducted on the reference lists when the titles appeared relevant.

Fig. 1
figure 1

Flow diagram of the scoping literature search

Full size image

Data collection

The following information was extracted independently by NK and NG from each paper: study design and purpose of the study, type of population studied, type of eating disorder and how the diagnosis was made, age, number of results and response rate included in the results of the study, and main results. (detailed in Tables 1, 2, 3 and 4).

Table 1 Articles concerning the school and general populations
Full size table
Table 2 Articles concerning sports professionals
Full size table
Table 3 Articles concerning general medical practice
Full size table
Table 4 Articles concerning health specialists
Full size table

Risk of bias

The risk of bias in individual studies was assessed using a quality rating provided by the Kmet form developed for quantitative analysis [34]. This form assesses the quality of studies on 11 different criteria in relation to the study design, methods, reporting of results, and study conclusions. Each study was given a score from 0 to 2 for each criterion, with a maximum total score of 22. The scores are shown in Appendix 1 in ESM.

Synthesis of results

The results were grouped according to four main types of study population: general and school populations, sports professionals, general medical practice, and health professionals.

Results

In all, 43 studies were included. These studies show some heterogeneity for the type of population, the countries where they were conducted, and the type of ED.

Types of population studied (Tables 1, 2, 3 and 4)

Among the 43 studies, the most widely represented populations were from schools (11/43), general medical practice (9/43), and sports (7/43), followed by other healthcare professionals (10/43): dental care professionals (4/43), gynaecologists (2/43), psychiatrists (1/43), gastroenterologists (1/43), and health professionals more generally including dieticians, psychologists, GPs, social workers, etc. (1/43) and medical residents (1/43) (Table 4). Finally, six studies in the general population were found (6/43) (Table 1).

The studies focusing on school populations (Table 1), including studies involving school personnel (4/43) and pupils (7/43) were the most numerous. This was followed by primary care in general medical practice and then by sports professionals (Table 2). Among primary care populations, the focus was mainly on general practitioners (7/43), and then on patients visiting GPs (2/43) (Table 3).

Source of the studies

Almost all the studies on general populations were conducted in Australia (5/6). Among the other studies (38/43), most of them were conducted in the USA (13/38), and the United Kingdom (11/38); the remaining 14 came from European countries (7/38), Canada (3/38) and Australia (4/38).

Type of ED understudy

The most widely studied type of ED in this sample of articles was AN (26/43), and BN came second (23/43). BED was very seldom explored (9/43). Some articles focused on ED symptoms (11/43). Six articles mentioned EDs without specifying which specific disorders they referred to.

Results in each type of population

In school and general populations

School populations

The approach to early detection were training staff about detection, and secondary prevention programmes directly applied to schools.

The first type of approach was to train school staff, and it seemed to enable EDs to be better recognised. One study presented ED training programmes for school staff and assessed a one-hour training programme delivered by an instructor who provided definitions, signs, and symptoms that were useful in identification, and talked about causes, effects and treatments. The staff was then asked to identify students they thought were at risk for an ED. The trained group identified significantly more students with an ED than the control group [37] (Table 1).

Two studies assessed school career orientation counsellors’ knowledge, their abilities and their degree of implication in the identification of students with AN or BN [35, 41]. These two studies assessed subjects' knowledge via a questionnaire: over twenty years, the counsellors’ knowledge improved for most items assessed in the questionnaire: recognition of symptoms of EDs, such as amenorrhea or binge episodes, had significantly improved [35]. The proportion of counsellors who thought they were able to identify an ED among students and help them if necessary did not however evolve over the 20 years (Table 1).

Finally, a study carried out on school staff more generally showed that the majority of staff felt helpless in the absence of a policy on EDs in their school; almost all of those who had not received any training thought that it would be useful for them [43] (Table 1).

The second type of approach was prevention programmes in schools. A study was conducted in 1997 in the United States in a university to assess a primary/secondary prevention programme: this programme involved an intervention by subjects having had EDs; the students filled in a self-assessment questionnaire three times over a duration of 12 weeks on their satisfaction or dissatisfaction with their current weight and their self-confidence. The researchers measured the impact of this primary prevention intervention on secondary prevention: 3 students with “a high risk of ED” out of 127, (corresponding to the first quartile of those who reported the most ED symptoms in the first survey), sought help after the programme had been completed. Proportionally, these students did not take part in this prevention programme any more than the others [36].

Some secondary prevention programmes have been implemented in American schools. Among them was the National Eating Disorder Screening Programme (NEDSP) [52]. It was initially implemented in 409 American universities in 1996. All the students were given an EAT-26 questionnaire to complete. 9069 students took part. Among them, 5787 saw a counsellor to talk about the results and particularly about the possible need for further exploration. These counsellors had been previously trained using slides and videos [53]. This programme was evaluated in universities. It was then implemented in high schools in 2000.

In universities, two years after the NEDSP, a sample of the students (see Table 1 for details) was randomly selected and interviewed on the phone. Only half of those who had been advised to see a doctor for a more in-depth examination had followed this advice by making a first appointment, and 39.4% stated they had sought care after the NEDSP. The authors concluded that school screening and educational programmes were good secondary prevention tools to identify and promote treatment for those presenting clinical symptoms [39]. Following the success of the NEDSP in universities, the experience was reproduced in 152 high schools in 34 American federal states in the early 2000s [54]. In four of these high schools, the assessment of this screening programme was conducted via a self-administered questionnaire two months after the programme implementation. Questionnaires from 1027 students were included in the analysis. More than half of them remembered their EAT-26 score. In all, 1/4 of the female adolescents interviewed and 1/5 of the male adolescents reported having shared their result with an adult. Of those who were advised to see someone, 11.5% reported they had been assessed by a health professional since they had been given their score. About 75% of the students interviewed recommended this screening programme to other schools. The members of the educational team who were interviewed thought that these explorations were important and they were willing to be involved [38].

Another intervention was run in the Netherlands in secondary schools: ex-patients who had had EDs came to take part in a one-hour session about EDs and how to identify them early. One-tenth of the students recognized ED symptoms in themselves and about a fifth recognized symptoms in other students. One-third of the students stated that they wanted to know more about the subject [40].

Finally, three Australian studies evaluated self-identification among students.

Two of them showed that nearly half of the adult students [42] and half of the secondary school adolescents [44] did not recognize themselves as having bulimic-type ED, illustrated in a case vignette, although the questionnaires showed high levels of symptoms. Those who recognized a problem with their eating significantly more often sought help than the non-recognizers [42, 44]. The third study results showed that a better self-identification of a body image problem significantly increased help-seeking among adolescents presenting an ED or an OSFED (Other specified feeding or eating disorder) or an UFED (unspecified feeding or eating disorder) [45].

In all cases, these training and/or screening programmes were well received by teachers and students and were considered useful.

General population

Six articles concerned secondary prevention of EDs in the general population.

One study evaluated the recognition of AN from a case vignette, and one-sixth of the participants correctly identified the diagnosis [48]. Another evaluated self-recognition of bulimic-type ED from a case-vignette of a woman presenting BN. Only half of the individuals (who present bulimic type symptoms on questionnaires) recognized having similar symptoms and those who self-identified significantly more often sought help from health professionals for eating and weight problems than the others [46].

Three studies evaluated training programmes in mental health first aid, among different types of university students and using different measures [47, 49, 50]. Two of them showed effectiveness in increasing knowledge about EDs and in improving recognition of BN symptoms [47, 49]. In the second program, “Should I say something”, at 3-months follow-up evaluation, nearly half of the students reported having encountered someone with a mental health problem or with an ED, and nearly all of them were able to provide assistance. They reported that when they suggested these people should seek help from professionals, half of them answered they would do so [49]. The third study, which was an internet dissemination of mental health first aid guidelines for EDs, showed that almost all participants found the documents useful, and a quarter of them tried to help someone after reading them [50].

Another type of study took place in Hamburg, Germany, from 2011 to 2014. It was intended to determine the effect of a public health campaign, the “Psychenet healthcare network campaign”, which focused on the time from the occurrence of AN symptoms to the first contact with the healthcare system and the initiation of specific treatment. The public health programme included short films on cinema screens, posters relating to these films, information on the Internet about AN, BN and BED, their treatment, a list of contacts for ambulatory or hospital care, and prevention in schools consisting in handing out leaflets listing the different treatment options for EDs. Two samples of female patients were recruited in treatment centres for AN in Hamburg: one group was created before the existence of this programme and another group afterwards. The results did not show any significant differences for the duration of illness before treatment, on average 3 years in both cases. As for the time it took to see a doctor, it was about 2 years after the onset of the illness in both situations. All in all, this public health campaign did not reduce the duration between symptom occurrence and specialist consultation/start of treatment [51].

Detection among sports populations (for detailed results see Table 2)

Different professionals from the professional world of sport were studied with respect to ED detection. They are indeed at the forefront when it comes to secondary prevention, as the prevalence of EDs (AN, BN and Eating disorders not otherwise specified—ED-NOS) is higher in sporting populations than in the general population [23].

These studies mainly focused on the sports professionals’ implication in early detection, their ability to detect EDs among individuals in sporting activities, their knowledge about EDs and their attitudes when they suspected an ED.

  • Most sports professionals feel involved in this kind of detection:

    • Almost all athletics trainers [55], thought it was their role to identify symptomatic individuals practising sport.

    • Among sports coaches from the American National Collegiate Athletic Association (NCAA) which organises sports programmes in many universities, female coaches were significantly more preoccupied by these questions among athletes. More than half of the coaches were interested in extra training on these pathologies, particularly on what to do once the symptoms were detected [56].

  • Fitness teachers. In one study, when shown a case scenario of AN in a young woman, over one-third of the coaches wondered about the ethical aspect of taking her into their club. This figure rose around two-thirds among those who had identified AN in this case scenario [57]. Similarly, in a Canadian study on a sample of fitness professionals, three-quarters of them stated they felt ethically obliged to intervene if they were confronted with a case of AN among the people they worked with [61].

    They were all interested in any recommendations on the subject [57].

  • Concerning their actual ability to detect an ED among people practising sports: one-third of the coaches from the NCAA stated they had referred an athlete with amenorrhoea to a doctor. They had identified between 2 and 3 symptomatic athletes on average during their careers [56]. One-third of Swedish elite coaches who were interviewed stated they were currently following an athlete with an ED [59]. Among athletics trainers, only a quarter were confident they could identify an ED in one of their athletes [55]. Indeed, more than a quarter of them reported having found out about an ED in an athlete after they had been prepared for competitions. Once the cases were identified, coaches relied mainly on resources from their athletics club and very seldom referred their athletes to GPs or specialists [56].

  • Knowledge about EDs, which is needed for detection by these different people, was incomplete and heterogeneous:

    • Among Norwegian fitness club coaches, less than one-third had accurate knowledge of ED signs and symptoms. The level of education was the only factor that significantly influenced their knowledge of EDs [60].

  • Thus, over one-third of sports coaches from the NCAA thought that amenorrhoea in an athlete was always normal [56]. Over half the sports coaches from fitness clubs, when presented with a typical clinical case of AN, considered that the person concerned was exercising too much, but they did not envisage the possibility of an ED, despite the fact that two-thirds of the sports coaches interviewed stated they had encountered club members in a similar situation to the case presented [57].

    As for elite coaches, two-thirds were only able to list one or two symptoms of BN [59]. Knowledge of the signs and symptoms required for identification was estimated at over two-thirds for coaches and for administrators of auxiliary dancers (majorettes, color guard, dance teams) on a true/false questionnaire [58]. One-third of Canadian professionals did not feel adequately prepared to manage this kind of situation [61].

  • Professionals’ attitude to potential cases of ED among sports people:

    • From 10% to one-third of fitness instructors in Canada and Norway thought they could talk to a young woman with AN about their worries concerning her state of health [57][57]. Another Canadian study showed that fitness professionals seemed inclined to intervene, as 59% of those who had suspected a case of AN stated they had done something about it. However, fewer than half thought that their intervention had some effect [61]. Around two-thirds of elite coaches said they would contact the athlete and tell her that they had observed symptoms of ED, but only 1/9 would refer them to a specialist [59].

Case detection by general practitioners (see Table 3 for detailed results)

General practitioners

These articles focused on different aspects of early detection: the rates of detection in general medical offices, factors influencing the rates of detection and the knowledge of EDs that is necessary for detection among GPs.

There was great variation in GP response rates in the studies on ED detection.

  • Rates of detection in medical consultation

The frequency of ED diagnoses established in general medical consultations was lower than what could be expected on the basis of epidemiological data. Certain authors [71] have estimated that the expected point prevalence of AN is 0.15%. On an active file of 2000 patients, one GP would therefore be confronted with 3 cases of AN. Flahavan et al. [64] estimated that 1–5% of the female population could be bulimic. Eighteen cases of BN on an active file of 2000 patients would therefore be expected on the basis of data from the “General Practitioner’s guide to eating disorders” [72].

Concerning AN, Irish GPs said they followed 1.34 patients on average [64]. It has been observed that for BN, the detection rate is well below the expected rate: in Ireland, GPs stated that they followed around one patient with BN on average [64]. These results were similar to a previous study conducted in the United States, which evidenced that 1/3 of GPs thought they had never met a patient with BN in their whole career and 2/3 declared not having any in their current active file [62].

In fact, health professionals do not always recognise these illnesses. In Philadelphia, 181 GPs watched videos of people with AN and were asked to give a diagnosis: 40% gave a correct answer, 20% gave several diagnoses including AN. However, 39% did not mention this possibility but instead thought of dysthyroidism, coeliac disease, gynaecological disorders, etc. For those who did give the diagnosis of AN, only 39% would refer the patients to a psychiatrist/psychologist [68]. A British study showed that only a third of GPs recognised 3 months of amenorrhea as a symptom of AN (in accordance with DSM-IV-R criteria in use at that time), and half of them mistakenly used a body mass index below 16 kg/m2 as a diagnostic criterion of AN [66].

Only one study focused on detection concerning BED; out of 2000 patients, GPs reported they followed on average 0.84 patients affected by this pathology [64]. We would expect more on the basis of epidemiological data [73].

Finally, in the United Kingdom, if all EDs are considered, the rate of referral by GPs to a specialised unit ranged from 0 to 4.14 per 1000 patients aged 65 years or under [63], which is lower than the expected figures.

  • Factors associated with ED detection by GPs

It has been shown that GPs who have already been confronted with cases of AN and BN outside their professional life significantly more often diagnose BN than others [62].

At the individual level, the referral rate has been reported to be higher in medical practices where there are several doctors, female GPs, younger doctors, those training other doctors, and those providing contraception; it was also higher when specialised centres were geographically close to the practice, and among doctors who had studied in the UK and not abroad. Finally, doctors who had Membership of the Royal College of General Practitioners (MRCGP) were more likely to refer patients with an ED [63].

Another study evidenced GPs’ anxiety towards weighing patients with eating disorder symptoms: the GPs did not know where to refer them, or how to convince them to receive treatment, and based themselves mainly on the patient’s weight to make their diagnosis [67].

Concerning GPs’ intentions to refer patients with EDs to a specialised unit, one study showed that intentions were significantly linked to subjective norms and cognitive attitudes, but not to the patient’s weight for instance, nor to the GP’s demographic characteristics [65].

Patients in general medical practice

As we mentioned above, results from studies on GPs showed that patients with EDs were less likely to be detected than could be expected on the basis of epidemiological data. One of the hypotheses investigated in the literature is that patients with EDs consult GPs less frequently than other patients.

However, two studies have shown that adult and adolescent patients with EDs see their GP significantly more frequently than other patients in the 5 years preceding diagnosis [69]. These results were confirmed among young people 14 years and under with AN [70].

The main reasons found for consultation were gastro-intestinal (diarrhoea and constipation), gynaecological or psychological symptoms [69].

Among younger patients with AN, the motives concerned weight, body image and eating. These preoccupations were generally not found among other patients [70].

Detection by health specialists (see Table 4 for detailed results)

These articles on early detection by health professionals mainly focused on the following points: their knowledge of EDs and assessments of e-learning programmes aiming to improve ED detection.

Dentists

Dentists and associated professionals play an important role in the secondary prevention of EDs, as they are the first to detect certain complications, especially with BN. They, therefore, have a crucial role in identifying and referring these patients.

However, concerning dental signs suggesting an ED, the dentists and the dental hygienists’ knowledge in the USA was generally only moderate; less than one-third of them obtained a score of 7/9 or more on a questionnaire on the subject [74].

Fewer than half of the dental hygienists actually looked for the oral symptoms of EDs and less than a fifth of them referred patients with an ED for specific treatment. Less than a tenth said that they were in contact with the patients’ GPs [75]. It is interesting to note that dental hygienists who had a greater perception of the seriousness of BN were twice as likely to assess the oral symptoms of EDs [75]. Concerning dentists, the situation was similar: those who recommended other forms of treatment for their patients’ ED were a minority (less than fifth). When dentists suspected an ED in one of their patients, in more than half of the cases, they did not mention it to the patient or to their family. However, for those who did share their suspicion, confirmation of their hypothesis was obtained in more than half the cases [77].

Norwegian dentists identified the media as their first source of information on EDs, followed by the Faculty of dentistry, and only in third position-specific lectures on the subject; moreover, more than half felt that they had not been given enough information on how to treat an ED [77]. An intervention in this respect was carried out in the UK. The objective was to assess an e-learning programme for dentists and dental hygienists, to improve secondary prevention of EDs via the provision of knowledge on the subject. The researchers concluded to a significant improvement in general knowledge, in knowledge of practice guidelines, in representations as to the specific role for their profession, in perceptions of the threat for the patients’ health, in the awareness of the benefit/risk ratio and in self-confidence in this field [76].

Gynaecologists, gastroenterologists, psychiatrists, residents and others

Consultations with gynaecologists, gastroenterologists and psychiatrists are favourable venues for ED detection. Indeed, amenorrhoea can precede weight loss in AN, BN can lead to menstrual irregularities despite a normal weight, and EDs can also lead to a risk of infertility. Gastrointestinal symptoms are also frequent among patients with an ED: abdominal distension, gastro-oesophageal reflux, constipation, abdominal pain, etc. [84].

The gynaecologists’ knowledge was fairly heterogeneous. Two-thirds were interested in training on the subject [79]. On the whole, they seemed to know about the psychopathology surrounding AN, but not so much for BN. Only 1/5 thought they could confidently diagnose an ED. In cases of oligomenorrhea or amenorrhea, 95% of gynaecologists looked for an associated ED. But more than a third did not assess the eating behaviours of patients with polycystic ovary syndrome [78], despite the fact they are more likely to have BN [85].

Concerning psychiatrists, a fairly heterogeneous level of knowledge about EDs was found: less than half of them linked amenorrhea to AN, one third mistakenly thought specific serotonin reuptake inhibitors were recommended for the treatment of AN, and 14.9% felt they were able to treat a patient with an ED; regarding bulimia, half identified body dysmorphic disorder (BDD) as a symptom of the illness [80].

Similarly, British gastroenterologists’ knowledge of AN and BN seemed incomplete. Even though the fear of putting on weight is a well-known criterion in AN, other criteria are not as well known: one-third thought that AN should be suspected only if the BMI was below 16 kg/m2. Half of them knew that BN is more frequent than AN. The issues relating to complications were better apprehended. Most of the professionals were not satisfied with the training they had received on this topic during their careers [81].

A study carried out on Canadian residents from different specialist fields showed that feeling comfortable with diagnosis increased significantly with each successive increase in hours of training on ED in children and adolescents; more than half were interested in extra training time for early identification and screening for child/adolescent EDs [82].

A recent Australian study showed how an e-learning programme proved useful to improve knowledge on EDs and the health professionals’ abilities and confidence in using this knowledge when treating patients with EDs [83].

Discussion

In terms of prevention, while quite a number of studies have been carried out on the primary prevention of EDs [32, 86], fewer studies have been conducted on early detection. It is however an important issue, as it has been demonstrated that early detection improves prognosis [4, 9]. A recent review showed that a shorter DUED could be related to a greater likelihood of remission [5]. Whereas the review by Le & al on primary prevention in 2017 described 112 studies between 2009 and 2015 [32], only 43 articles on early detection were found up to 2021. They mainly focused on AN and BN, and only nine articles explored the case of BED. This could be explained by the fact that the diagnostic criteria changed over time: BED diagnosis first appeared as a proposal in DSM IV-R in 1994 and as a defined diagnosis in the DSM 5 (2013).

To sum up, the area and aims of the studies on early detection were very heterogeneous. Three main areas are the focus of interventions aiming to improve early detection: healthcare (19/43) including general medical care (9/43), schools (11/43), and sport (7/43). The aims were to evaluate (1) among professionals: knowledge about EDs, attitudes towards identification and management of EDs, training programs on EDs, detection rates for EDs and patterns of consultations before ED diagnosis (30/43) and (2) among students or the general population: evaluation of prevention programmes (7/43), and evaluation of knowledge and self-recognition (6/43).

We will discuss here the main findings, firstly the question of the level of knowledge about EDs among professionals and how to improve it, and secondly the results of the main interventional studies, the role of GPs in early detection, and finally proposals for future studies.

Level of knowledge about EDs among professionals and how to improve it

This was the main domain investigated by the studies reviewed.

Whatever their profession, the professionals liable to be involved in early detection had only partial knowledge of the symptoms to be recognised for the identification of EDs [35, 41, 56, 59, 60, 66,67,68, 74, 77,78,79,80,81, 83]. It should also be noted that the majority of the studies were based on surveys, generating recruitment biases. Individuals willing to participate may already be familiar with these pathologies.

For residents, the greater the number of training hours on EDs, the more confident they felt in their ability to diagnose [82]. There is a need to improve the levels of knowledge among healthcare professionals (initial training and continuing education). Some authors suggest that in initial training curricula, specific training modules in the area of early identification of EDs should be included [82]. But this means finding windows in health professional training to include new content in already crowded schedules. An advantage of e-training interventions is that they can be used as a resource or tool to supplement existing curricula for health professionals such as dentists who require both didactic and skill-based training [76]. E-learning programs about EDs can also be interesting for continuing education among healthcare professionals. These programs are a cost-effective method of knowledge and skill-building. They can reach a wide audience at relatively low cost, enabling the distribution of learning materials to promote mental health literacy among health professionals and, in turn, improve the diagnosis and treatment of patients with EDs in the health system [77, 83]. Thus, several proposals exist but need to be evaluated.

International guidelines provide information on how to diagnose EDs. However, only 3.8% of British GPs reported using local published guidelines or protocols, and none of them use national guidelines [87]. Some authors, therefore, think information needs to be provided in a form that is readily accessible when patients with ED present in primary care or in psychiatry, rather than in the form of guidelines that are not used because it is not easy to find the information needed [66, 80]. This is why international guidelines have developed short documents focused on the question of early detection not only for professionals but also for families and patients [88,89,90,91,92].

Assessment of professional knowledge is only an intermediate step in the detection process, but patient-centred studies focusing on early detection are scarce.

Results of the main interventional studies

There were very few interventional studies about the improvement of early detection, whatever the population.

One important area for early detection is the school environment, as school age is the most frequent period of ED onset. To improve early detection, one team proposed to train school personnel [37] on the identification of students at risk for ED, how to approach them and where to refer them. They obtained an improvement in the detection of students with these pathologies [37]. Some authors have developed interventions centered directly on students; one study showed that primary prevention is not a suitable method to improve early detection [36]. Two different teams evaluated one and the same prevention program in high schools and in universities. The results showed that school screening combined with educational programs (NEDSP) were successful secondary prevention tools to identify and promote treatment for those presenting clinical symptoms of EDs in universities [39], but the effects on early detection were less conclusive in the high school NEDSP program evaluation [38]. Another program proposed to train students by asking ex-patients to explain the first symptoms of ED; these sessions enabled some students to recognize symptoms of ED in themselves or in other students [40]. Indeed, studies have shown that self-identification of a body image problem [45] or a problem with eating [44] increased help-seeking among adolescent students. Two prevention programs focused on university students, training them for mental health first aid and they showed an improvement in ED recognition in the general population [47, 49]. Finally, a public health campaign did not reduce the time-lapse between symptom occurrence and specialist consultation/start of treatment for AN [51].

Role of primary care in early detection

GPs are key professionals in early detection, as patients with EDs seem to have seen their GP significantly more frequently than other patients in the five years preceding ED diagnosis [69, 70]. However, GPs very rarely mention ED diagnoses in their patient population compared to what could be expected from the epidemiological data [62, 63]. Reasons for the low detection and referral rates for AN remain poorly understood [68]. As previously mentioned, there may be a lack of training [93], and poor knowledge of EDs could also be explained by the relative rarity of these cases in primary care. Demographic variables can also impact the detection rate for EDs, for instance gender and ethnicity of the physician and ethnicity of the patient [68]: female physicians are more likely than male physicians to diagnose AN [63], and white/Caucasian physicians are more likely to correctly diagnose a vignette of a patient with AN, especially if the patient in the vignette is also white/Caucasian [68].

Some results also show that BN is clearly less well detected in primary care than AN. This may be due to the fact that as BMI is lower among patients with AN, this pathology is more easily detected during a physical examination. Clinicians may also miss cases of BN because of the belief that it is a rare condition in their population [66]. Furthermore, BN is more likely to be experienced as a shameful illness, so that patients do not spontaneously talk about the symptoms [94]. Stigma and shame are two of the most prominent barriers to help-seeking for EDs [95], as is lack of insight into the disorder [46, 96]. De-stigmatising BN could be a lever for its detection. Indeed, in the United Kingdom, some authors have described the Diana effect [97]: shortly after public revelations on Princess Diana’s BN, the number of BN cases increased. This is probably because people who had BN were able to talk about their symptoms and seek help [98]. Further to this, people suffering from bulimic-type ED are more likely to recognize their pathology when they self-induce vomiting; thus better information about non-purging type BN needs to be incorporated into prevention programs [42, 46]. A better dissemination of mental health literacy on eating disorders [99] among students and the general population could help to improve self-identification and help-seeking [44,45,46, 50]. Mental health literacy has been defined as “knowledge and beliefs about mental disorders which aid their recognition, management or prevention”[100].

Proposals for future studies

To increase early detection, research should seek to identify methods to reduce barriers to help-seeking (stigma, shame, denial, lack of ED literacy, negative attitudes towards treatment) [95]. Almost all the studies about self-recognition were conducted in Australia; thus it would be interesting to replicate them in other countries and to include males subjects [42].

Research is needed on young people's living environments, particularly school and university, and also sport centers. Finally, it would be interesting to evaluate proposals that appear in the literature.

For example, future studies should evaluate the impact of training for school personnel by previously trained school nurses in countries where this is possible [37]; or detection of EDs by periodically checking weight [101]. Other studies should try to identify means whereby students suspected of having an ED could be more strongly motivated to seek evaluation by a health professional when they are detected [38, 39]. Mental health first aid approaches as proposed in Australia [102] or sessions delivered to students by ex-patients [40] could be evaluated in controlled trials, with longer follow-up periods [47, 49].

Similarly, the prevalence of AN, BN and ED-NOS is higher among people practising sports than in the general population [23]. Thus, research should be developed to evaluate the implementation of specific recommendations for the field of sport [103] and sport at the university [55], after training sessions for professionals [56]. Guidelines of this type have been developed, for example in Australia by Fitness Australia [104]. Fitness centers should also have lists of local clinicians and therapists to counsel clients with a suspicion of ED [60]. Another suggestion is to make information brochures available for fitness professionals and clients providing information on EDs and local treatment centers [61] and to evaluate their impact.

It also seems important to publish the recommendations on EDs for each profession and each speciality concerned, as is already the case with gynaecologists in the United States [105] or in France [106], and with dentists in France [107].

It is also important to widen the health professional focus in training for early detection, as some people who could be in a position to detect EDs were not included in the published studies, for instance, endocrinologists, pharmacists, or patients’ relatives (in the general population).

It is also worth noting that apart from a few studies on school students [36, 38,39,40, 42, 44, 45] and on the general population [46,47,48,49,50,51], most of the articles reviewed here were centred on people in charge of interventions. Assessment of their knowledge is only an intermediate phase in the detection process. It would be interesting to assess certain aspects related to the patients or the people concerned themselves.

Limitations

The field is very heterogeneous and there are few studies on each different theme or each population, leading to a poor level of evidence. In addition, none of the studies in the review met all the quality criteria defined by the Kmet form [34]. But as mentioned earlier, early detection is a key point in treatment effectiveness [5] and to our knowledge, this review is the first to be performed on this question, and to suggest the fields that should be investigated in future studies.

What is already known on this subject?

International guidelines mention the need for early detection. But the means to achieve this have not been detailed. Literature reviews on primary prevention exist but not on secondary prevention.

“What your study adds”

This review summarises all studies on the early detection of EDs. Different types of the population can be focused on to improve early detection, including health professionals, students and school personnel, sports professionals, and the general population. There are several means to improve early detection, which are described and discussed in this review.

Conclusion

Early detection of EDs can be carried out in all sectors of the lives of young people and adults, particularly by adults in the close circle, whether in school, sport or healthcare.

It should be based on a prevention policy that targets information dissemination and training in order to improve detection. Finally, this policy should be implemented alongside a policy for the organisation of care by way of a clear, accessible care trajectory and patient-centred assessments.