Scientific papers

Recent trends in Adult ADHD and Autism assessment referrals Introduction § 1 ADHD and autism are currently the most recognised neurodevelopmental conditions (NDCs) in adult psychiatry with adult prevalence rates of 2-4% and 1% (1) respectively, making them more common than schizophrenia. § 2 Their importance for mainstream adult mental health services has only been recognised recently - those with ADHD or Autism have significantly increased risks of suicide, other major mental disorders, and shortened life expectancy (2). § 3 Diagnostic expertise and services are still developing, with British diagnostic prevalence rates for both trailing significantly behind expected prevalence rates (3). § 4 Unfortunately, many of Britain’s fledgling adult ADHD and Autism assessment services have been overwhelmed recently because of unexpected increases in demand (4). § 5 A cogent response requires quantitative data. Assessment demand has already been partially quantified for Autism (5) but not for ADHD. § 6 Indirect evidence suggests a larger rise in demand for ADHD compared to Autism: NHS England has needed to respond rapidly to the crisis by forming an ADHD Programme Clinical Reference Group (6); § 7 many clinicians from regional Autism and ADHD services (within the purview of the Neurodevelopmental Disorders Special Interest Group (NDD SIG) of the Royal College of Psychiatrists) have reported unexpected increases in ADHD referrals further exceeding capacity (4); § 8 service saturation has caused some NHS ADHD services (7) and ‘Right to Choose partners’ (8) to either pause their ADHD waiting lists, prioritise high risk individuals (e.g. 9), or quote lengthening waiting times; § 9 the ADHD Foundation (a neurodiversity charity providing pre-diagnostic screening) quote a 400% increase in demand for ADHD assessment since 2020 (10); § 10 ADHD assessment in the UK private sector has recently expanded inordinately (6); § 11 finally, the US and UK shortage of ADHD medication infer that demand increases were unforeseen (11, 12). § 12 However, there is little quantitative data proving increased ADHD demand. A cogent response also needs to identify factors driving demand. § 13 These are currently undefined, although there are at least two contemporarily relevant candidates- the influence of social media and of Covid-19 pandemic restrictions. § 14 In support of the first, front-line clinical observation suggests that social media may have increased awareness of ADHD: § 15 a member of the author group whilst working in primary care noted many seeking ADHD assessment had done so after using social media, particularly TikTok, suggesting that social media may have driven referrals. § 16 In support of the second, the unexpected increase occurred around the time of the Covid pandemic, suggesting pandemic specific factors such as covid societal restrictions may have been important. § 17 The first aim of this study was to quantify adult ADHD assessment referral rates between 2019 to 2023 and to compare this with adult Autism referral rates from the same sample areas. § 18 Further, we investigated if ADHD medication prescribing trends also increased unexpectedly, indirectly marking an increase in the rate of those being diagnosed with ADHD. § 19 The second aim was to explore whether changes in referral rates were associated with changes in public interest in ADHD and/ or related to COVID-19 restrictions. § 20 We also explored if social media increased the specificity of those being referred for ADHD assessment. Methods § 21 Anonymised referral data that would be freely and publicly available through a relevant freedom of information request was used to measure referral rates in various regions of the UK. § 22 In no cases were any specific patient identifiable material used. Freely available Google search data and data related to Covid restrictions were also gathered and compared with ADHD referral data. 1. ADHD and Autism assessment referral rates § 23 Anonymised routinely and systematically collected monthly referral data between October 2018 and April 2023 from six UK regional adult autism and six ADHD services covering the same geographical population in a part of Scotland and across areas in the Southeast, the Southwest, and the Northeast of England were analysed. § 24 For each service, the average monthly referral rate for 6-month aliquots from beginning April to end September, and beginning October to end March, normalised to per 100,000 people aged 18 and above covered by that service was calculated. § 25 As referral rates were broadly similar across condition specific services, mean monthly referral rates and 95% confidence intervals for each condition specific service for each time aliquot were calculated using XLStat (13). § 26 2. Public interest in autism and ADHD Volumetric data was obtained from the ‘Keywords Everywhere’ algorithm (14) which measured UK-specific Google searches for the terms ‘adhd’ and ‘autism’. § 24 The average monthly search for each 6-month aliquot, was standardised to per 10,000 people aged 18 to 65 using the UK population figures from the Office for National Statistics (15). § 25 3. Birth gender and age at referral. One service provided composite ADHD and Autism referral data on birth gender and average age at referral. § 26 A gender ratio indicator was calculated by subtracting the proportion of women from the proportion of men. A figure of 0 represents a M:F ratio of 1:1, -1 representing a Male: Female ratio of 2:1, and +1 representing a M:F ratio of 1:2. § 27 4. Prescribed ADHD medication. Numbers and gender distribution of people in England receiving prescribed ADHD medication from 2015 to 2022 was ascertained using freely available data from the NHSBSA (16). § 28 5. Stringency of Governmental Covid measures and public interest in ADHD. Publicly available daily Stringency Index (SI) data from the Oxford Coronavirus Government Response Tracker (OCGRT) project (17) was used. § 29 The SI is an index of a government’s daily Covid-19 stringency measures. The peak SI and the duration for which the SI was continuously 50 or greater (50 chosen arbitrarily) were noted. § 30 The Average Sustained Intensity of Measures (ASIM50) was calculated by totalling the daily SIs for the period when they were continuously 50 or over, divided by the number of days. § 31 The lower and upper limits of the ASIM50 lay between 50 and 100. These parameters were examined for 4 countries- the UK, the US, Sweden, and India, chosen because of their differing government Covid pandemic measures. § 32 All four countries allowed free access to TikTok during the study period. Relative interest search data from Google Trends (18) for the search terms ‘ADHD’ for the UK, USA, Sweden, and India over the past 5 years were identified. Volumetric data was not available for India or Sweden. § 33 Google Trends produces relative search figures (indicating relative interest) for a specified search term for the chosen country, assigning ‘100’ to the maximum number of searches in a week within the specified period, with the country’s other weekly figures being scaled accordingly. § 34 For each country, average relative interest for ‘ADHD’ over the 6-month aliquots described above were calculated. We quantified the magnitude of increase in Google searches as: 6. Exploring if internet media had a specific effect in identifying those with possible ADHD § 35 The ratio of volumetric UK Google searches to UK referral rates for ADHD was calculated. Referrals to assessment services are made if they exceed specific clinical thresholds. § 36 If thresholds remain unchanged and referral rates increase, a reduction in this ratio could occur if a larger proportion of people presenting to primary care exceeded this threshold. § 37 This would indirectly suggest that internet searches may have helped identify a greater proportion of appropriate referrals. § 38 If there was no change in the ratio in the face of increasing referral rates, this could suggest that internet searches generally raised awareness without increasing specificity. § 39 Results 1. Referral rates (figure 1) The sampled services covered around 10% of the British population (approx. 6.7 million people). § 40 Two services provided composite ADHD and Autism referral data, four providing separate ADHD and Autism referral data. § 41 Despite different geographical locations and service configurations, there was a consistent pattern and rate of referral across all services, all experiencing unexpected increases from July 2020. § 42 From July 2020 to January 2023 (2.5 years) ADHD and Autism services saw a x2.9 combined increase in referrals (17 vs 50 referrals/100000/ month). § 43 This was mainly attributable to increased ADHD referrals which tripled from 11 to 33 /100000/month, an average increase of 80% per annum. § 44 In comparison Autism referrals increased x1.72 (8.5 to 14.6 /100000/month). (a) Combined ADHD or Autism assessment referrals across all regional services. § 45 (b) ADHD referrals across 4 regional ADHD services. (c) Autism referrals across 4 regional Autism services § 46 (d) UK-wide Google searches for keywords ‘ADHD’ or ‘Autism’ (e) UK-wide Google searches for keyword ‘ADHD’ (f) UK-wide Google searches for keyword ‘Autism’ § 47 2. UK Public interest in ADHD and Autism (figure 1). From July 2019 to January 2023, Google searches for ADHD or autism increased x2.6 (from 33 to 85/ month). § 48 ‘ADHD’ increased x3.8 (from 13 to 50 searches/ month/ 10000). Both ADHD searches and referrals increased unexpectedly from July 2020. § 49 The magnitude of increase in both from July 2020 were very similar (figure 1 (d) to (f)). 3. Birth gender and age of those referred. § 50 One service provided birth gender data and age of those referred for ADHD or Autism assessment (figure 2). § 51 Pre-pandemic, more males than females were referred. This changed gradually such that by January 2022, more females than males were referred. § 52 The Male: Female ratio gradually switched from January 2019 (2:1 Male to Female) to July 2023 (1:1.5 Male to Female). § 53 There was no change in the average age of those being referred before July 2020 (mean= 31.5 years) compared to after (mean= 31.4 years). § 54 Figure 2a: number of referrals for ADHD or Autism assessment by gender from January 2019 to July 2023. § 55 Figure 2b: gender ratio indicator for ADHD and Autism referrals between January 2019 and July 2023. An indicator of -1 equates to a M:F ratio of 2:1, 0 equates to a ratio of 1:1 and 1 equates to a ratio of 1:2. § 56 4. ADHD prescribing data. The average annual rate of increase in those prescribed ADHD medication doubled after 2020 (figure 3). § 57 Before 2020, prescriptions rose by 10.4%/ annum, whereas after 2020 prescriptions rose by 21%/ annum. During this time, ADHD referral rates remained relatively static prior to 2020 but rose by 80%/ annum after 2020 § 58 Post-covid, there was a 146% increase in prescriptions in the 30-34 age group compared to a 28% increase in the 10-14 age group. § 59 The 25-44 age group correlates to the largest users of TikTok (19), encompassing 49.8% of users. § 60 5. Internet searches for ‘ADHD’ and Covid stringency measures for four countries All four countries showed increased searches from July 2020. § 61 On average, ADHD searches rose x2.4 between July 2020 and January 2023 (range from x1.7 to x3.8). Figure 4a- relative Google searches for ADHD across 4 countries § 62 Figure 4b- relationship between size of increase in Google searches and ASIM50. All 4 countries imposed restrictions greater than 50 on the SI in the 2nd to 3rd week of March 2020. § 63 The UK had the highest ASIM50 and Sweden the lowest (table 1) Figure 4b demonstrates the relationship between the magnitude of increase in interest in ADHD from July 2020 to January 2023, and the ASIM50. § 64 Increasing ASIM50 was associated with increasing searches for ADHD. Neither the duration of the ASIM 50, nor the peak SI were associated with changes in searches. § 65 6. Ratio of internet searches to referrals The ratio of ADHD searches to ADHD referrals in the UK was 13:1 and changed little over 4 years, suggesting internet media may have had a generalised awareness raising effect. § 66 Discussion Findings To the best of our knowledge, this paper is the first to provide evidence that social media and pandemic social restrictions may have had a direct major impact on mental health service demand. § 67 Referral rates for ADHD increased from mid-2020 for reasons transcending service parameters. § 68 In absolute terms, demand for ADHD assessments was greater than for Autism from the beginning of the sample period (January 2019), and accelerated disproportionately in the Covid and post-Covid period. § 69 Average ADHD referral rates rose unexpectedly by a factor of 3 from 11 to 33/month/ 100000. This roughly matched rises in public interest in ADHD (x 3.8). § 70 ADHD prescriptions also rose unexpectedly after 2020 but were limited, possibly representing the delayed effects of increased ADHD referrals and diagnosis, the rise potentially limited by service capacity. § 71 Preliminary data also suggests a significant shift from predominantly males to females being referred, also unexpected given the historically accepted gender distributions for ADHD. § 72 This may have started pre-pandemic, with gender ratio reversing from end of 2021. This shift parallels recent prescribing data showing a marked rise in ADHD medication prescriptions for women (16). § 73 Furthermore, it was observed that the greater the degree of imposed sustained COVID restrictions (ASIM50), the greater the increase in the public’s interest in ADHD. § 74 TikTok’s usage increased extraordinarily during 2020, either coincidentally or as an unforeseen consequence of Covid related social restrictions. § 75 Downloads of the TikTok application peaked in the first half of 2020 and it was the most downloaded non-gaming application in 2021 and 2022 (21). § 76 ADHD has been described as “massive on TikTok” (22), #ADHD being viewed around 1.19 billion times a month from 6/22 to 11/23 (23), equivalent to 20 views per month per 100 of the world’s 15-65 population. § 77 It was not possible to obtain detailed data on the number of active TikTok users or the numbers exposed to ADHD on TikTok by year and country. § 78 However, it can be argued that this may not be the most informative or accurate metric: unlike other social media, TikTok’s personalised recommendations algorithm includes content that may not be directly related to the person’s interests (24). § 79 Instead, a more representative measure of a person’s sustained interest may be internet searches initiated by individuals themselves. § 80 Therefore, Google searches was seen as a better proxy measure of public interest, Google being the most frequently used internet search engine worldwide. § 81 The second line of enquiry was whether Covid social restrictions had any bearing on referrals, the Covid pandemic being the most prominent event of 2020. § 82 Whilst Covid may be associated with subsequent neuropsychiatric symptoms (20), these are unlikely to represent ADHD which is characterised by lifelong traits causing significant dysfunction and, by definition, would predate Covid. § 83 It is more likely that referral rates were influenced by indirect or social consequences, or by coincidental factors. § 84 For this, the ASIM50 was derived from the OCGRT’s stringency index as a proxy measure of the sustained intensity of restrictions applied to a country’s population. § 85 As it was not possible to obtain reliable referral data for other countries, Google search data was used given its possible association with referral rates. § 86 From mid-2020, ADHD searches increased across all four countries in keeping with the start of Covid restrictions. § 87 Possible interpretation It seems that several partially independent events came into play around the time of the Covid pandemic to lead to the large unpredicted increase in ADHD assessment. § 88 Firstly, we know that adult ADHD was significantly under identified compared to expected prevalence rates prior to the pandemic, particularly in women (2). § 89 Secondly, like most of the world, the UK population underwent significant social restrictions from early 2020. § 90 However, unlike previous pandemics, people had easy internet access allowing more people to use electronic media for social contact, as well as for information. § 91 Thirdly, perhaps coincidentally or because of lockdown restrictions, TikTok’s use ‘exploded’ around this time. § 92 TikTok’s unique algorithm potentially exposed the public to material that they would not necessarily seek out, leading to a world-wide audience watching videos labelled as ADHD, without ADHD being their primary interest. § 93 This raised awareness of ADHD, possibly more for ADHD in women. This then led to more people seeking information through internet searches, which then led to more people seeking assessment. § 94 The data suggests that this increased awareness generally: it did not change the ratio of searches to referrals. § 95 The possible correlation between the ASIM50 and internet searches also suggests that more severe sustained restrictions were associated with greater use of social media, greater exposure to ADHD related material and then subsequent internet searches. § 96 A possible implication is that demand for adult ADHD assessments also increased in other countries proportionate to the increase in public interest, although this remains to be proven. § 97 Whilst it is possible that lockdown measures had a disproportionately negative impact on those with unidentified ADHD, it is not possible to conclude this with certainty based on our findings. § 98 Several factors are therefore suggested to have coincided to lead to the unprecedented rise in ADHD referrals: pre-existent diagnostic under-identification; §99 social restrictions increasing social media use; TikTok’s unique algorithm; increased general awareness; easy access to information. Social media and mental health conditions. § 100 The effects of TikTok and other social media on awareness raising and identification of health conditions, and their impact on services is yet to be determined. § 101 Preliminary evidence suggests that younger adults may be turning to social media to self-diagnose or self-identify (25) the main perceived benefit being support and validation (26). § 102 However, social media’s diagnostic accuracy is highly variable. § 103 Yeung et al (27) found that around half of the ‘ADHD’ TikTok videos were misleading, and Olvera et al (28) found the depictions of tics on TikTok differed significantly from typical tic disorders. § 104 This raises legitimate concerns that increased assessment demand may be driven by those who do not meet diagnostic criteria. § 105 The finding that the searches to referrals ratio did not change suggests that this was not the case and supports the argument that social media had a general, awareness raising effect. § 106 Limitations Our service sample encompassed a broad range of socioeconomic and demographic backgrounds and service configurations. § 107 Despite this variability, referral data and patterns of change were sufficiently similar across all services to allow us to describe them using summary statistics. § 108 Furthermore, the additional advantage was that the sampled services covered a heterogeneous 10% sample of the British population, improving the generalisability of the findings. § 109 This study necessarily relied heavily on derived or proxy measures of public interest, social media use and social restriction. § 110 It is also recognised that these findings can only ascribe associations rather than establishing causality. § 111 Despite these limitations, the data is strongly compatible with the notion that social media had a significant and enduring effect on ADHD referral rates beginning early during the Covid pandemic. § 112 Implications for the future Demand for ADHD assessments was greater than for Autism from 2019 and increased disproportionately from 2020. § 113 Social media and easily available electronic information appear to have already had a major unforeseen and enduring impact on adult ADHD assessment demand. § 114 It is known that public interest can drive demand, this being the essence of advertising. § 115 However, the current findings are novel and important for two reasons: they establish a significant ‘bottom up’ drive in demand, the public itself having identified ADHD as being important independent of health professionals’ views, § 116 and they illustrate how rapidly, unpredictably, and profoundly health service demand can increase with electronic social media, ready access to information and the appropriate social milieu. § 117 The most pressing task now is to develop an appropriate medium to long-term health service response. § 118 Understandably, individual services have responded with short term measures generally prioritising those perceived to be at greatest clinical risk, as no service has sufficient capacity to absorb an unplanned 300% increase in demand. § 119 The current study provides timely information that may support future decision making and service development directions at a time where there is an emergent focus on the commissioning demands for ADHD service in England. § 120 Whilst several options have already been described, each with benefits and drawbacks (4), stakeholders now need to collaborate to answer the fundamental questions of ‘who benefits most from a medical diagnosis in dimensional disorders,’ § 121 ‘who benefits most from a psychiatric/psychological approach to ADHD’ and ‘who benefits most from medication’, the lower cut-offs for each being unknown. § 122 These are key questions for a ‘realistic medicine’ approach (29), which may help guide the optimal use of health service time, expertise, and resource. § 123 It is envisaged that this data will lead to further research studies into these key unknowns, these being the cornerstones to developing a thoughtful strategic response. § 124 Acknowledgement We would like to thank Gordon Love, for his editorial guidance and proof-reading efforts on the grammar and structure of the paper.