Sage Journals: Discover world-class research

Abstract

Recent developments in artificial intelligence (AI) and machine learning raise the possibility of screening and early diagnosis for neurodegenerative diseases, using 3D scans of the retina. The eventual value of such screening will depend not only on scientific metrics such as specificity and sensitivity but, critically, also on public attitudes and uptake. Differential screening rates for various screening programmes in England indicate that multiple factors influence uptake. In this narrative literature review, some of these potential factors are explored in relation to predicting uptake of an early screening tool for neurodegenerative diseases using AI. These include: awareness of the disease, perceived risk, social influence, the use of AI, previous screening experience, socioeconomic status, health literacy, uncontrollable mortality risk, and demographic factors. The review finds the strongest and most consistent predictors to be ethnicity, social influence, the use of AI, and previous screening experience. Furthermore, it is likely that factors also interact to predict the uptake of such a tool. However, further experimental work is needed both to validate these predictions and explore interactions between the significant predictors.

Keywords

Health screening predictors of screening health psychology public health artificial intelligence public attitudes Parkinson’s screening neurodegenerative diseases

Introduction

Neurodegenerative diseases are a major and growing cause of disability. For example, the number of people diagnosed with Parkinson’s in the UK is expected to almost double between 2018 and 2065,¹ with an even faster predicted increase in Alzheimer’s.² Currently in Parkinson’s for example, most of the dopaminergic neurones and some essential functioning of the brain have already been lost before diagnosis,³ indicating that earlier diagnosis is crucial in beginning effective treatment. The retina is part of the central nervous system, directly visible from the outside, and can now be imaged non-invasively in near-microscopic detail by optical coherence tomography (OCT), a relatively inexpensive method that is increasingly available to detect eye disease in high-street opticians. Mounting evidence suggests that neurodegenerative diseases are associated with changes in retinal structure long before physical symptoms occur,⁴ offering the possibility of early diagnosis via retinal imaging. However, these changes are subtle and barely detectable by humans, requiring considerable training and experience to detect. Advances in artificial intelligence (AI) and machine learning offer a possibility for earlier diagnosis, through training AI algorithms on large datasets to recognise these subtle changes. Therefore, this raises the possibility that people attending a routine eye test could be screened to detect neurodegenerative diseases very early in the disease process.

Alongside the practical question as to how such screening would be implemented is the critical question of whether the public would accept such screening. Since the technology described above offers the potential that such a screening tool may be developed, it is important to explore and understand these issues before it is approved for use. Potentially, if public acceptance is very low, screening might not be appropriate at all. More realistically, an exploration into the factors that influence acceptance may reveal concerns which can be addressed by appropriate action, such as better engagement with particular communities. This literature review will explore the potential influencing factors using existing screening initiatives as a guide.

Predictive factors for AI-enabled screening

It is important to acknowledge that much of what may be said about the likelihood of screening uptake for neurodegenerative diseases depends on the analysis of past screening uptake behaviour. However, as the number of screening programmes in the UK is small and uptake varies vastly, this analysis must be supplemented by analysis of public health studies exploring the predictors of other preventative health behaviours such as vaccination. Thus, in the absence of a validated screening tool to explore predictors through the collection of primary data, the possible predictors of AI-enabled screening for neurodegenerative diseases may only be explored through consideration of a wide range of literature. The possible predictive factors discussed in this review fall into three main groups:

a) Properties of the disease itself and early diagnosis of it (e.g. awareness, perceived benefit of early diagnosis, perceived risk and susceptibility).

b) Properties of the screening procedure (e.g. perceived accuracy, inconvenience, pain, embarrassment and time).

c) Properties of the individual (e.g. their perceived level of vulnerability, perceived control over their own life, trust of health authorities or authority in general, previous medical experiences).

Clearly, several of these could apply and also interact. For example, a negative experience with a particular screening method (a) might mean that method is met with resistance when encountered again (b). In this review, we will assess the available evidence relating to each of these groups in turn, in the context of AI-enabled screening for neurodegenerative disease via OCT imaging. It is important to note that given the interactions aforementioned, the grouping will necessarily be somewhat arbitrary.

The remaining discussion of this review identifies several factors which might influence take-up of screening for neurodegenerative diseases if this became possible in future, based on existing literature of health screening. Namely, these are the influence of participants’ age, sex/gender, perceived mortality risk, source of the diagnosis (AI vs human) and the stated benefits of early diagnosis of, and personal proximity to neurodegenerative disease.

Properties of the disease itself and its early diagnosis

Multiple studies have found that being affected by a particular disease vicariously, through the diagnosis of a family member, predicts uptake of early screening tools. For example, relatives are more likely to attend mammography,^5,6 cervical^6–8 and colorectal^9,10 cancer screening, and lack of family history has been cited as a barrier to attendance.¹¹ A further relevant finding is that most relatives of those with Parkinson’s would accept genetic testing if offered,¹² despite the relatively low genetic risk.¹³ There are several reasons why being affected by a neurodegenerative disease might predict uptake of an early screening tool.

Increased understanding of neurodegenerative diseases and benefits of early diagnosis

One possibility is that relatives are more accepting of early diagnostic methods because they have increased awareness of the illness, which subsequently predicts uptake. Lack of awareness of colorectal cancer¹⁴ in comparison to other, more highly publicised cancers¹⁵ may help to explain the observation that uptake and coverage for cervical and breast cancer in those eligible lies between 70% and 75% in England,^16,17 whilst uptake of bowel screening is far lower at 57.7%.¹⁷ However, an alternate suggestion is the relatively high time investment compared to other screening types, as the screening via faecal occult blood test requires submission of three samples. Indeed, implementation of the faecal immunochemical test (FIT) test⁷ which requires one sample increases uptake by 8.5%.⁶ Nonetheless, uptake still does not reach that of cervical, breast, or diabetic eye screening, retaining the possibility that a lack of awareness is the largest determinant of this decreased uptake.

Indeed, lack of knowledge around why it is important to attend screening and the available treatment options for the disease being screened for has been identified as one of the main barriers to diabetic eye,¹⁸ colorectal,¹⁴ and breast^5,19 screening and increased perception of seriousness was one of the main predictors of breast and cervical screening in the US.⁶ Additionally, social media campaigns have been demonstrated to significantly increase knowledge,²⁰ intention²¹ and subsequent screening attendance by 12.9%.²² Other successful knowledge-based interventions include an intervention comparing the discomfort of cervical screening to the potential consequences of not attending screening, which increased attendance by 15%,²³ and an intervention of diabetes education, which significantly increased attendance for diabetic eye screening.²⁴ However, many campaigns include other elements such as providing social support²⁰ and information about the procedure,²² thus it is difficult to tease their effects apart. Also, the success of interventions to increase awareness may be dependent on there being benefits to early diagnosis.^25,26 Although lifestyle changes²⁷ or medication²⁸ may be effective for some, there is currently no cure or definitive effective treatment for neurodegenerative diseases such as Parkinson’s. Therefore, the perceived benefits of early diagnosis are likely to better predict screening for neurodegenerative diseases than awareness alone.

Perceived risk of neurodegenerative disease

Being related or close to someone with a neurodegenerative disease may increase its salience and therefore perceived risk for the disease; additionally, the former might induce perception of increased genetic risk. However, the evidence on perceived risk as a predictor of screening is inconsistent, as a meta-analysis of 18 longitudinal studies found that perceived susceptibility to an illness showed a minimal overall correlation of −0.06 with preventative health behaviours, and perceived benefit of the behaviour emerged as a much stronger predictor.²⁹ This may be explicable by the finding that the relationship between perceived risk and cervical screening was mediated by cancer worry.³⁰ Since high cancer worry is rare even in high risk populations,³¹ the relationship is only detectable in large samples. As relatives of those with Alzheimer’s report high worry towards developing the disease,^32,33 this may drive a relationship between perceived risk and screening uptake for neurodegenerative diseases.

It is important to note, though, that perceived risk does not necessarily reflect actual risk. Although in relatively older women (age 40–86; 312 women) for osteoporosis, breast cancer and heart disease perceived risk has been shown to decrease with age, actual risk for these diseases increases with age.³⁴ This decline in perceived risk might explain the greater likelihood of younger women (aged 35–50; 581 women) versus older women (50 and over) to attend recommended mammography screening.³⁵

Social influence

Social influence, through wider sources such as the media and healthcare providers, is a well-established predictor of screening uptake. Discussion with and encouragement from not only relatives,^10,36 but also physicians^10,34,37,38 and wider social groups^10,36,39 have been shown to strongly predict colorectal,¹⁰ breast^34,37,39 and prostate^36,38 cancer screening. It is likely that physicians, specifically primary healthcare providers, exert the largest influence on screening attendance; for example, screening for colorectal cancer in England increased by 6% when endorsed by a letter from a local GP and by a further 11% when the letter was signed by the GP.⁴⁰ Further, a review and meta-analysis of 69 interventions found a modest but consistent effect of a GP’s signature on invitation letters.⁴¹ The consistent trust in physicians and subsequent influence in patient’s screening decisions has led to the conclusion that their relationship is the most important predictor of screening attendance.³⁷

Social influence may also represent the mechanism between the relationship between family history and enthusiasm towards screening, since family history is a significant predictor of discussion around screening.⁴² However, it is most likely that these factors also act independently, as there is evidence that the effect of discussion with friends, family and healthcare providers on screening interacts with age. Whilst younger women were found to rely on these sources for breast cancer screening, middle aged women were more likely to utilise mass media.⁴³ Interestingly, an intervention study of African American women found that whilst age, marital status, discussing with friends, and believing early detection led to a cure were all positive predictors of mammography attendance, family history of breast cancer was no longer a significant predictor.³⁹ However, discussion with friends and family did not mediate the relationship between family history and prostate cancer screening, whilst number of acquaintances with prostate cancer did.³⁶ As all participants were African American, the influence of number of acquaintances may depend on ethnicity, which will be further discussed as a predictive factor below.

In summary, history of disease and discussions with friends and family are proposed to predict higher uptake of screening for neurodegenerative diseases, due to an increased perceived risk mediated by worry, and social interaction, respectively. Additionally, perceived benefits of screening and endorsement from a local GP are likely to be independent predictors.

Properties of the screening procedure

AI versus human agent

A number of studies have found that a human healthcare provider is generally preferred over AI,^44,45 demonstrated across a range of prevention, treatment and diagnosis scenarios.⁴⁴ Moreover, this preference is most prominent for decision making within healthcare, including diagnosis.⁴⁶ Mainly, AI is only accepted for these purposes providing a physician is present to supervise or for a second check,^47,48 indicating that the public do not support the unsupervised use of AI for screening and diagnosis.⁴⁷ Studies have found different reasonings for this decreased trust in diagnosis,⁴⁹ such as perceptions that AI cares less than a human doctor,⁵⁰ a perceived lack of control over AI being used for diagnosis,⁴⁹ and perception of lower accuracy.⁵¹ However, even if the accuracy of AI were proven to be as accurate as physicians’, participants still felt more confident in a human doctor over AI.⁵² Alternatively, a preference for a human doctor was found to be mediated by the view that AI does not account for the uniqueness of each patient and their health.⁴⁴

However, this preference runs deeper than conscious reasoning. A comparison of explicit and implicit attitudes towards AI found that whilst only 6% reported a negative attitude to AI, 77% showed a negative implicit attitude on the implicit associations test (IAT).⁵³ This test measured sub-conscious biases to find that most participants more readily conflated AI with ‘bad’ rather than ‘good’, compared to humans. Additionally, a study involving fMRI found that whilst participants reported preferring AI when the conversation was personalised rather than generic, the areas of the brain associated with prosociality were only activated when engaging in the same conversation with a human doctor.⁵⁴ This selectiveness towards feeling a social and trusting connection indicates an inbuilt preference towards humans as an evolutionary mechanism.⁵⁴

While the tendency to distrust AI is general, a number of other factors may interact with it. Evidence for the influence of demographic factors such as age and sex/gender is mixed,^52,55 and may be dependent on other factors such as country.⁵⁵ Additional predictors of acceptance towards AI include education,⁴⁸ direct previous experience with AI,^46,53 and personal history⁵² and severity⁴⁶ of the disease the AI aims to diagnose. It is likely that the use of AI will tend to lower the acceptance of screening for neurodegenerative diseases, although this may change as public awareness of and experience with AI becomes more widespread.

Experience of screening using OCT

Although no AI algorithms for diagnosing neurodegenerative disease from OCT images have yet been approved for use in the UK, OCT is already used to diagnose ophthalmic diseases such as glaucoma,^56,57 diabetic retinopathy⁵⁸ and age-related macular degeneration.^58,59 A relevant finding is, therefore, that whilst there are multiple factors that predict previous screening behaviour, only previous screening behaviour remains a significant predictor of current screening.^60,61 Furthermore, a study of 2657 women found that important predictors of initial breast cancer screening such as family history and proximity to someone with breast cancer were no longer significant in predicting attendance for the second and third time.⁶²

One explanation is that attending screening eliminates any worries or fear around the screening procedure. In support, familiarity with the screening procedure is a firm predictor of mammography screening⁵ and campaigns that aim to provide information about the procedure have successfully improved screening attendance.¹⁷ A beneficial feature of OCT imaging is that it is painless, convenient and involves no bodily exposure. Screening for eye health in diabetic patients is similar to the procedure of OCT imaging and shows a much higher uptake than other screening programmes in the UK, at 82.7%.¹⁷ Indeed, beliefs that diabetic eye disease could be seen with an eye examination was a significant, positive predictor of self-reported adherence to screening. Furthermore, pain, convenience and bodily exposure have all been demonstrated to predict screening uptake.^63–67 Therefore, previous experience with OCT should eliminate any fears around pain or embarrassment.

Home sampling for cervical and colorectal cancer greatly increases uptake,⁴¹ potentially indicating that factors such as embarrassment were the main barriers. However, having the home sampling kit delivered rather than needing to collect it from a GP increases uptake still further.^11,14,41,68 This suggests that in fact it is the convenience of home screening which is the major advantage. Despite this, it remains reasonable to conclude that previous experience of OCT scanning within an optometrist or optician would predict a high uptake of screening tool for neurodegenerative diseases that utilise OCT, due to the consistent predictor of past screening behaviour.

Properties of the individual

Socioeconomic status and uncontrollable mortality risk

One consistent finding in the literature is that those of lower socioeconomic status (SES) are more likely to: suffer from adverse health outcomes such as obesity (in women⁶⁹), engage less in healthy behaviours and more in harmful health behaviours,⁷⁰ and have a lower life expectancy.⁷¹ For example, average life expectancy for men in the UK was found to range from 73 for those in very unskilled manual occupations, compared to over 80 for those in the highest professional roles.⁷² For each intermediate occupational class, there is a step decrease in life expectancy,⁷² suggesting an underlying gradient⁷³ rather than a ‘cliff edge’ effect,⁷⁴ wherein there is a sudden decrease in life expectancy below a certain SES threshold. More specifically, one review of inequalities due to SES noted that the differences also exist in preventative healthcare and can sometimes be substantial.⁷⁵ For example, on the rollout of screening for colon cancer in the UK, there was a 26% difference in uptake between those at the bottom fifth and top fifth SES,⁷⁶ despite the existence of free universal healthcare in since 1948.⁷⁷ Many studies have found a similar relation between some measure of SES and probability of some protective health behaviour – for example, between women’s occupation and flu vaccination,⁷⁸ education and cervical screening,⁷⁹ and area SES and both follow-up diagnostic tests⁸⁰ and eye tests⁸¹ – although the effect size is rarely as substantial as for colon cancer.⁷⁶ Furthermore, this effect is also present across other developed countries in Europe,^82,83 and is especially linked to cancer screening.^82,83 Considering the most important driver of the relationship between SES and preventative health, education remained an independent predictor of cervical screening even when income, class and occupation were added into the model.⁷⁹ Overall, a review of the effect of education concludes that it is at least as strong a predictor as income, and that the two interact.⁸⁴

One possible reason for the link between education and screening is that highly educated people have higher health literacy, that is are better able to comprehend health information and thus make informed decisions about their health.⁸⁵ Although a review of 10 articles of mostly self-reported cancer screening found relatively inconsistent findings,⁸⁶ a longitudinal study with over 3000 participants found a 10% difference in attendance for colorectal cancer screening between those with acceptable versus with limited health literacy, even when controlling for age and gender.⁸⁷ Furthermore, in a sample of 2668 adults, health literacy mediated the relationships between education and: flu vaccination, dental check-ups and mammography screening.⁸⁸ Although participants were over 65 and all variables were measured cross-sectionally, this does indicate a mechanism for the strong relationship between education and preventative health behaviour including screening.

However, those of lower SES also experience higher exposure to environmental risks such as air pollution,^89,90 poor housing quality,^89,91 and infections such as COVID-19,⁹² a finding that differences in health literacy does not account for. Whilst stress and a lack of resources⁷³ have been suggested as the link between adverse environment and investment in preventative health, an underlying evolutionary mechanism is posited by the uncontrollable mortality risk hypothesis.⁹³ The hypothesis states that when an individual is faced with a number of threats to mortality that are out of their control- as is the case in those of low SES- they will invest less in behaviours within their control. This is because if an individual’s environment threatens the longevity of their life, there is little point in investing energy and resources into preventative health behaviours. In a series of experiments, Pepper and Nettle⁹⁴ demonstrate that manipulating the perceived controllability of life expectancy significantly predicts whether a fruit or chocolate prize is chosen, irrespective of whether participants were primed with long or short life spans. This indicates the effect is due to the uncontrollable aspect rather than a general mortality threat. Moreover, in an online study of 600 North Americans, Pepper and Nettle⁹⁵ found uncontrollable mortality risk to fully mediate the relationship between subjective SES and health investment.

Since the uncontrollable mortality risk hypothesis posits that lack of control results in prioritisation of short-term outcomes over long-term, it makes sense that it would apply to preventative health behaviours including screening. Indeed, actual⁹⁶ and perceived⁹⁷ mortality risk significantly predict health screening behaviour and interest in it,⁹⁶ in studies where the influence of controllability was not investigated. Also, health optimism was found to interact with mortality salience, such that women high in health optimism were significantly more likely to intend to engage in breast self-examination when primed with a mortality salience cue than women who were not.⁹⁸ Although this study⁹⁸ purportedly investigated health optimism, the scale items related to an individual’s ability to control their health outcomes and thus were more reflective of controllable mortality risk. Unlike optimism, which is viewed as a trait because it is relatively stable over time,⁹⁹ uncontrollable mortality risk is dependent on the environment and changes in response to cues.⁹⁴ The relationship between perceived controllability of risk and preventative health behaviours for COVID-19 reveals complexities of the interaction, as fear related to the COVID-19 pandemic was found to interact with perceived control over it, so that in those of low perceived control, fear increased engagement in preventative health behaviours both generally and relating to the pandemic, whereas for those of high perceived control, increased fear decreased engagement in preventative health behaviours and had little effect on COVID-19 preventative health behaviours.¹⁰⁰ This finding may appear to contradict the mortality risk hypothesis, however the included scale measured perceived control over contracting and being ill with COVID-19, rather than mortality. In fact, there are no existing studies on COVID-19 that have specifically addressed uncontrollable mortality risk. Thus, it is reasonable to predict that low SES and perceived uncontrollable mortality risk will negatively predict AI-enabled screening for neurodegenerative disease via OCT imaging.

Given the heterogeneity of findings related to demographic factors and preventative health behaviours, findings for sex/gender, age and ethnicity are next presented in table format, summarising the directions of their predictions in screening for neurodegenerative disease.

Age

As shown in Table 1, the relationship between age and health behaviour depends on the health behaviour. Age negatively predicts cancer screening^35,79,87,101 but positively predicts vaccination,^102,103 especially in older populations¹⁰⁴ or for COVID-19,^102,103 uptake of screening for eye conditions relating to eye health¹⁰⁵ and attendance of NHS health checks.¹⁰⁶ A large international review of flu vaccination intentions and behaviour found the prediction from age to be highly inconsistent,¹⁰⁴ although the review was international and so these polarised findings may be due to differences in vaccination initiatives and incentives. Additionally, age may also interact with factors such as SES³⁵ to predict screening.

Table 1.

Summary of studies reporting a relationship between age and willingness to accept screening/vaccination.

Study	Country	Sample size	Age range	Outcome measure	Direction of relationship
Gerend et al.¹⁰¹	US	312	40–86	Perceived susceptibility to breast cancer	Negative
Champion³⁵	US	581	50+	Compliance with mammography screening guidelines screening	Negative
Sabates and Feinstein⁷⁹	UK	4275	22–65	Uptake of cervical screening	Negative
Kobayashi et al.⁸⁷	UK	3087	50+	Uptake of colorectal cancer screening	Negative
Cochrane et al.¹⁰⁶	UK	10,483	40+	Uptake of NHS health checks for vascular conditions- offered to identified people at risk	Positive
Lawrenson et al.¹⁰⁵	UK	291,296	12+	Uptake of diabetic eye screening	Positive
Schmid et al.¹⁰⁴	International	Review of 470 articles	ranged	Intentions and uptake of flu vaccination	Mixed (more for positive, studies looking at elderly populations were all positive)
Robertson et al.¹⁰²	UK	12,035	16+	COVID-19 vaccination hesitancy	Older age = less vaccine hesitant
Woolf et al.¹⁰³	UK	11,584	M = 45	COVID-19 vaccination hesitancy	Older age = less vaccine hesitant

Sex/gender

It is important to note that the difference in UK screening uptake of colorectal cancer and cervical and breast cancer may be due to a sex/gender difference, namely that women are more likely to attend screening. However, the global review by Schmid et al.¹⁰⁴ found polarised results for the relationship between sex and flu vaccination, with very few studies finding no significant differences; possibly due to publication bias. Namely, the authors concluded that gender was a key predictor of flu vaccination in elderly, healthcare professional and general populations, but for those with chronic diseases. Being female was more frequently cited as a barrier, although many studies also cited it as a facilitator. Conversely, there is consistent evidence that being male predicts COVID-19 vaccination in both US¹⁰⁷ and UK^102,103,107 samples. Men are 41% more likely to intend to receive the vaccine on average,¹⁰⁸ an effect which is fully mediated by perception of others as intending to be vaccinated, perceived control over vaccination, and less anticipation of negative feelings such as fear.¹⁰⁹ For other preventative healthcare, large UK studies found females to be more likely to attend colorectal screening,^87,110 males to have a higher uptake of NHS health checks for those at risk of vascular diseases,¹⁰⁶ and no effect of gender on the uptake of free eye examinations.¹¹¹ As shown in Table 2, the effect of sex is largely dependent on the type of preventative healthcare, and thus it is uncertain whether or how sex might affect screening for neurodegenerative diseases.

Table 2.

Summary of studies reporting a relationship between sex/gender and willingness to accept screening.

Study	Country	Sample size	Age range	Screening type	Direction
Shickle and Farragher¹¹¹	UK	17,680	Groups from under 16 to over 65	Free eye examination	None
Cochrane et al.¹⁰⁶	UK	10,483	40–74	Health checks offered to at risk Ps	Males more likely
Kobayashi et al.⁸⁷	UK	3087	50+	Colorectal cancer screening	Females more likely
Wardle et al.¹¹⁰	UK	ALL people eligible for screening across England	60–74	Colorectal cancer screening	Females more likely (not sig tested)
Zintel et al.¹⁰⁸	International	141,550 across 66 studies	Primarily adult samples	COVID-19 vaccine intentions	Males more likely
Sieverding et al.¹⁰⁹	UK and Germany	1428	M = 41	COVID-19 vaccine intentions	Males more likely

Ethnicity

Ethnicity as a predictor of preventative health behaviours is more consistent (Table 3). In Western countries, Black, Asian and other ethnic minority participants are less likely than White participants to attend flu vaccination,^88,104 prostate cancer¹¹² and colorectal cancer¹¹² screening. The same pattern is observed for cervical screening in the UK,^79,113 even after controlling for SES.¹¹³ One literature review concluded that ethnicity is an independent predictor of preventative health behaviours, and therefore should not be collapsed into an overarching variable of SES.¹¹² In the US and UK, participants who are an ethnicity other than White are also more hesitant towards COVID-19 vaccination than White participants,^{102,103,107,114} but this only translates to reduced attendance in the US and not the UK.¹⁰⁷ Hesitancy is highest in Black participants but is also increased in Asian and Hispanic participants.^102,107 Breast cancer screening is not predicted by ethnicity in the US⁸⁸ and UK,¹¹³ but is in Canada,¹¹⁵ although the review of studies in Canada does not quantify this effect. As the effect of ethnicity is not accounted for by SES, a potential alternative explanation is health literacy. For example, as for the relationship with education, health literacy was found to mediate the relationship between ethnicity and flu vaccination, but not for mammography screening or dental check-ups.⁸⁸ Indeed, COVID-19 vaccine hesitancy was found to partly be caused by a lack of knowledge,¹⁰⁷ indicating health literacy to be a mediator of ethnicity for vaccines, but not screening. The lack of ethnicity diversity in vaccine study samples and lack of trust have also been cited as causes of COVID-19 vaccine hesitancy.¹⁰³ In the case of screening, distrust may be particularly acute when healthcare is carried out by AI, given the longstanding built in biases in AI towards ethnic majorities due to sampling bias.¹¹⁶ By nature, ethnic minorities are less represented when building AI algorithms, so the AI becomes less accurate for them. Previous examples have found schizophrenia to be over diagnosed,¹¹⁷ and prediabetes and diabetes to be underdiagnosed¹¹⁸ in ethnic minorities. Provided people of ethnic minorities are aware of these previous issues, this may lead to distrust and hesitance towards screening using AI. In summary, those of Black ethnicity are the least likely to take part in all preventative care aside from mammography, for reasons such as distrust which is arguably justified given that non-White people still experience discrimination within healthcare,^119,120 resulting in poorer treatment of ethnic minorities.¹²⁰

Table 3.

Summary of studies reporting a relationship between ethnicity and willingness to accept screening and vaccination.

Study	Country	Sample size	Type	Direction	Size
Sabates and Feinstein⁷⁹	UK	4275	Attendance of cervical screening	Asian Ps less than White Ps	8.4% less
Moser et al.¹¹³	UK	3185	self-report of breast and cervical screening	No sig difference for breast, cervical: White women more likely	OR = 2.2 compared to other ethnicities
Schmid et al.¹⁰⁴	International	470 articles	Flu vaccination intentions and behaviour	Inconsistent- In all included UK and most US studies, being White predicted vaccination.	No meta-analysis
Bennett et al.⁸⁸	US	2668	Flu vaccination, dental checkups and mammography screening	White more likely to attend but not for mammography. No sig difference to Latino Ps.	Ethnicity → vaccination relationship mediated by health literacy
Chiu¹¹²	Mostly US	55 studies	Cervical, breast, prostate, colorectal	Low uptake of BAME groups	Due to knowledge, SES and health literacy. Low quality
Hanson et al.¹¹⁵	Canada	52 studies	Mammography	Ethnic minority = less uptake	One of strongest barriers, but the study does not quantify this
Willis et al.¹¹⁴	US	1205	COVID vaccination hesitancy	White Ps less likely to show hesitancy than African/ American Ps	African American Ps 2.5+ × more likely to show hesitancy, low trust = 11+ × likely to show hesitancy
Nguyen et al.¹⁰⁷	US and UK	87,000+ and 1.2 million+, respectively	COVID vaccination hesitancy and uptake	BAME groups more hesitant than white Ps, Black Ps the most hesitant, only translated to uptake in the US	OR = 3.15 for Black participants, 1.42 for Hispanic, 1.34 for Asian participants. Black Ps were less knowledgeable, more likely to work on front line and have been infected with covid before
Woolf et al.¹⁰³	UK	11,584 healthcare workers in UK	COVID vaccination hesitancy	BAME groups more likely to be hesitant	OR = Black Caribbean: 3.37 Black African: 2.05 compared to white British Ps. Hesitancy caused by lack of trust, ethnic diversity in vaccine studies and confusing info (qual data)
Robertson et al.¹⁰²	UK	12,000+	COVID vaccination hesitancy	Higher in BAME groups	OR = 13.42 in Black and 2.54 in Pakistani/Bangladeshi groups compared to white Ps. Also higher in lower education levels

Ps is used as shorthand for participants.

Conclusion

From the findings of this literature review, the most important positive predictors of the acceptability and uptake of AI-enabled screening for neurodegenerative disease via OCT imaging, in the UK, are likely to be:

1) Knowing someone/possessing a relative with a neurodegenerative disease (mediated by social influence).

2) Perceived risk (mediated by worry).

3) Perceived benefits of screening.

4) The use of a human agent rather than AI.

5) Previous experience of receiving an OCT retinal scan.

6) Education (mediated by health literacy).

7) High SES (mediated by uncontrollable mortality risk).

8) Being an ethnic majority (i.e. White in the UK) compared to an ethnic minority.

The above factors are not an exhaustive list, but represent those for which the evidence discussed above is most consistent. Future qualitative and quantitative research would benefit from specifically exploring the acceptability of AI-enabled screening for neurodegenerative disease via OCT imaging. This literature review provides a basis to explore the above factors in relation to uptake of such a tool.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: ‘This report is independent research funded by the National Institute for Health Research and NHSX (Artificial Intelligence, “OCTAHEDRON : Optical Coherence Tomography Automated Heuristics for Early Diagnosis via Retina in Ophthalmology and Neurology”, AI_AWARD01976).

Significance for public health

This study highlights the importance of investigating the factors that determine uptake of a new screening tool alongside the development of the tool itself. Further research around the potential influencing factors named in this study will help inform interventions to increase uptake when screening for neurodegenerative diseases using OCT and artificial intelligence becomes available. Furthermore, this narrative review provides valuable insights for predicting health screening behaviour in general, and how predictors might change depending on the individual, the disease that is being screened for, and the nature of the screening procedure.

ORCID iDs

Beth AB Nichol

Anya C Hurlbert

Jenny CA Read

References

Parkinson’s UK. The Incidence and prevalence of Parksinon’s in the UK. Results from the Clinical Practice Research Datalink Summary Report. 2018.

Wittenberg

Barraza-Araiza

, et al Projections of older people with dementia and costs of dementia care in the United Kingdom, 2019–2040. London: London School of Economics, 2019.

Blesa

Trigo-Damas

Dileone

, et al Compensatory mechanisms in Parkinson’s disease: circuits adaptations and role in disease modification. Exp Neurol 2017; 298: 148–161.

La Morgia

Di Vito

Carelli

, et al Patterns of retinal ganglion cell damage in neurodegenerative disorders: parvocellular vs magnocellular degeneration in optical coherence tomography studies. Front Neurol 2017; 8: 710.

Subramanian

Oranye

Masri

, et al Breast cancer knowledge and screening behaviour among women with a positive family history: a cross sectional study. Asian Pac J Cancer Prev 2013; 14(11): 6783–6790.

Yamal

Atkinson

, et al Predictors of breast and cervical screening in Vietnamese women in Harris County, Houston, Texas. Cancer Nurs 2005; 28(2): 119–129; quiz 130.

Chan

Choi

Wong

, et al Examining the cervical screening behaviour of women aged 50 or above and its predicting factors: a population-based survey. Int J Environ Res Public Health 2016; 13(12): 1195.

Erbil

Tezcan

Gür

, et al Factors affecting cervical screening among Turkish women. 2010.

Young

McGloin

Zittleman

, et al Predictors of colorectal screening in rural Colorado: testing to prevent colon cancer in the high plains research network. J Rural Health 2007; 23(3): 238–245.

10.

Madlensky

Esplen

Gallinger

, et al Relatives of colorectal cancer patients: factors associated with screening behavior. Am J Prev Med 2003; 25(3): 187–194.

11.

Chorley

Marlow

Forster

, et al Experiences of cervical screening and barriers to participation in the context of an organised programme: a systematic review and thematic synthesis. Psychooncology 2017; 26(2): 161–172.

12.

Falcone

Wood

Xie

, et al Genetic testing and Parkinson disease: assessment of patient knowledge, attitudes, and interest. J Genet Couns 2011; 20(4): 384–395.

13.

Tan

E-K

Jankovic

Genetic testing in Parkinson disease: promises and pitfalls. Arch Neurol 2006; 63(9): 1232–1237.

14.

Honein-AbouHaidar

Kastner

Vuong

, et al Systematic review and meta-study synthesis of qualitative studies evaluating facilitators and barriers to participation in colorectal cancer screening. Cancer Epidemiol Biomarkers Prev 2016; 25(6): 907–917.

15.

Bong

McCool

Chinese peoples’ perceptions of colorectal cancer screening: a New Zealand perspective. NZ Med J 2011; 124(1331): 29–38.

16.

NHS Digital. Breast Screening Programme: England, 2017-18, https://files.digital.nhs.uk/60/77DCCC/breast-screening-programme-eng-2017-18-report.pdf (2019).

17.

NHS

England.

Report of the independent review of adult screening programmes in England, https://www.england.nhs.uk/wp-content/uploads/2019/02/report-of-the-independent-review-of-adult-screening-programme-in-england.pdf (2019).

18.

Prothero

Lawrenson

Cartwright

, et al Barriers and enablers to diabetic eye screening attendance: an interview study with young adults with type 1 diabetes. Diabetic Med 2022; 39: e14751.

19.

Stoll

CRT

Roberts

Cheng

M-R

, et al Barriers to mammography among inadequately screened women. Health Educ Behav 2015; 42(1): 8–15.

20.

Plackett

Kaushal

Kassianos

, et al Use of social media to promote cancer screening and early diagnosis: scoping review. J Med Internet Res 2020; 22(11): e21582.

21.

Klippert

Schaper

Using Facebook to communicate mammography messages to rural audiences. Public Health Nurs 2019; 36(2): 164–171.

22.

GTF. Using Facebook to improve attendance at screening appointments. 2018.

23.

Morrell

Perez

Hardy

, et al Outcomes from a mass media campaign to promote cervical screening in NSW, Australia. J Epidemiol Community Health 2010; 64(9): 777–783.

24.

Zou

Y-H

Cui

S-S

, et al Predictors for attending annual eye screening for diabetic retinopathy amongst patients with diabetes in an urban community of Beijing. Int J Ophthalmol 2017; 10(7): 1144.

25.

Beeker

Kraft

Southwell

, et al Colorectal cancer screening in older men and women: qualitative research findings and implications for intervention. J Community Health 2000; 25(3): 263–278.

26.

Palmer

Midgette

Dankwa

Colorectal cancer screening and African Americans: findings from a qualitative study. Cancer Control 2008; 15(1): 72–79.

27.

Ahlskog

JE.

Does vigorous exercise have a neuroprotective effect in Parkinson disease?

Neurology 2011; 77(3): 288–294.

28.

Bogetofte

Alamyar

Blaabjerg

, et al Levodopa therapy for Parkinson’s disease: history, current status and perspectives. CNS Neurol Disord Drug Targets 2020; 19(8): 572–583.

29.

Carpenter

CJ.

A meta-analysis of the effectiveness of health belief model variables in predicting behavior. Health Commun 2010; 25(8): 661–669.

30.

Zhao

Nan

The influence of absolute and comparative risk perceptions on cervical cancer screening and the mediating role of cancer worry. J Health Commun 2016; 21(1): 100–108.

31.

Hay

Buckley

Ostroff

JS.

The role of cancer worry in cancer screening: a theoretical and empirical review of the literature. Psychooncology 2005; 14(7): 517–534.

32.

Cutler

SJ.

Worries about getting Alzheimer’s: who’s concerned?

Am J Alzheimers Dis Other Demen 2015; 30(6): 591–598.

33.

Navab

Negarandeh

Peyrovi

Lived experiences of Iranian family member caregivers of persons with Alzheimer’s disease: caring as ‘captured in the whirlpool of time’. J Clin Nurs 2012; 21(7–8): 1078–1086.

34.

Nekhlyudov

Ross-Degnan

Fletcher

SW.

Beliefs and expectations of women under 50 years old regarding screening mammography: a qualitative study. J Gen Intern Med 2003; 18(3): 182–189.

35.

Champion

Relationship of age to mammography compliance. Cancer 1994; 74(1 Suppl): 329–335.

36.

McDowell

Occhipinti

Chambers

SK.

The influence of family history on cognitive heuristics, risk perceptions, and prostate cancer screening behavior. Health Psychol 2013; 32(11): 1158–1169.

37.

Roberts

Birch

NJ.

A comparison of breast cancer secondary prevention activities and satisfaction with access and communication issues in women 50 and over. Prev Med 2001; 32(4): 348–358.

38.

Cormier

Reid

Kwan

, et al Screening behavior in brothers and sons of men with prostate cancer. Urol J 2003; 169(5): 1715–1719.

39.

Husaini

Sherkat

Bragg

, et al Predictors of breast cancer screening in a panel study of African American women. Women Health 2001; 34(3): 35–51.

40.

Hewitson

Ward

Heneghan

, et al Primary care endorsement letter and a patient leaflet to improve participation in colorectal cancer screening: results of a factorial randomised trial. Br J Cancer 2011; 105(4): 475–480.

41.

Camilloni

Ferroni

Cendales

, et al Methods to increase participation in organised screening programs: a systematic review. BMC Public Health 2013; 13(1): 464–516.

42.

Carter-Harris

Tan

Salloum

, et al Patient-provider discussions about lung cancer screening pre- and post-guidelines: Health Information National Trends Survey (HINTS). Patient Educ Couns 2016; 99(11): 1772–1777.

43.

Ogata Jones

Denham

Springston

. Effects of mass and interpersonal communication on breast cancer screening: advancing agenda-setting theory in health contexts. J Appl Commun Res 2006; 34(1): 94–113.

44.

Longoni

Bonezzi

Morewedge

CK.

Resistance to medical artificial intelligence. J Consum Res 2019; 46(4): 629–650.

45.

Lennartz

Dratsch

Zopfs

, et al Use and control of artificial intelligence in patients across the medical workflow: single-center questionnaire study of patient perspectives. J Med Internet Res 2021; 23(2): e24221.

46.

Scott

Carter

Coiera

Exploring stakeholder attitudes towards AI in clinical practice. BMJ Health Care Inform 2021; 28(1): e100450.

47.

Ongena

Yakar

Haan

, et al Artificial intelligence in screening mammography: a population survey of women’s preferences. J Am Coll Radiol 2021; 18(1): 79–86.

48.

Jonmarker

Strand

Brandberg

, et al

The future of breast cancer screening: what do participants in a breast cancer screening program think about automation using artificial intelligence?

Acta Radiol Open 2019; 8(12): 2058460119880315.

49.

Juravle

Boudouraki

Terziyska

, et al Trust in artificial intelligence for medical diagnoses. Progress in Brain Research 2020; 253: 263–282.

50.

Yokoi

Eguchi

Fujita

, et al Artificial intelligence is trusted less than a doctor in medical treatment decisions: influence of perceived care and value similarity. Int J Hum Comput Interact 2021; 37(10): 981–990.

51.

Nelson

Pérez-Chada

Creadore

, et al Patient perspectives on the use of artificial intelligence for skin cancer screening: a qualitative study. JAMA Dermatol 2020; 156(5): 501–512.

52.

Jutzi

Krieghoff-Henning

Holland-Letz

, et al Artificial intelligence in skin cancer diagnostics: the patients’ perspective. Front Med 2020; 7: 233.

53.

Fietta

Zecchinato

Di Stasi

, et al Dissociation between users’ explicit and implicit attitudes toward artificial intelligence: an experimental study. IEEE Trans Hum Mach Syst. 2021; 52: 481–489.

54.

Yun

Lee

Kim

DH.

Behavioral and neural evidence on consumer responses to human doctors and medical artificial intelligence. Psychol Market 2021; 38(4): 610–625.

55.

Sindermann

Yang

Elhai

, et al Acceptance and fear of Artificial Intelligence: associations with personality in a German and a Chinese sample. Discov Psychol 2022; 2(1): 1–12.

56.

Ojima

Tanabe

Hangai

, et al Measurement of retinal nerve fiber layer thickness and macular volume for glaucoma detection using optical coherence tomography. Jpn J Ophthalmol 2007; 51(3): 197–203.

57.

Garas

Vargha

Holló

Reproducibility of retinal nerve fiber layer and macular thickness measurement with the RTVue-100 optical coherence tomograph. Ophthalmology 2010; 117(4): 738–746.

58.

Adhi

Duker

JS.

Optical coherence tomography–current and future applications. Curr Opin Ophthalmol 2013; 24(3): 213–221.

59.

Schmidt-Erfurth

Klimscha

Waldstein

, et al A view of the current and future role of optical coherence tomography in the management of age-related macular degeneration. Eye 2017; 31(1): 26–44.

60.

Rutter

DR.

Attendance and reattendance for breast cancer screening: a prospective 3-year test of the theory of planned behaviour. Br J Health Psychol 2000; 5(1): 1–13.

61.

Sweetman

Watson

Norman

, et al Feasibility of familial PSA screening: psychosocial issues and screening adherence. Br J Cancer 2006; 94(4): 507–512.

62.

Drossaert

CHC

Boer

Seydel

. Prospective study on the determinants of repeat attendance and attendance patterns in breast cancer screening using the theory of planned behaviour. Health Psychol 2003; 18(5): 551–565.

63.

Consedine

Krivoshekova

Harris

CR.

Bodily embarrassment and judgment concern as separable factors in the measurement of medical embarrassment: Psychometric development and links to treatment-seeking outcomes. Br J Health Psychol 2007; 12(Pt 3): 439–462.

64.

Aro

Koning

HJD

Absetz

, et al Two distinct groups of non-attenders in an organized mammography screening program. Breast Cancer Res Treat 2001; 70(2): 145–153.

65.

Koo

Leong

Ching

, et al Knowledge of, attitudes toward, and barriers to participation of colorectal cancer screening tests in the Asia-Pacific region: a multicenter study. Gastrointest Endosc 2012; 76(1): 126–135.

66.

Matthias

Imperiale

TF.

A risk prediction tool for colorectal cancer screening: a qualitative study of patient and provider facilitators and barriers. BMC Fam Pract 2020; 21(1): 43–48.

67.

Ghanouni

Nuttall

Wardle

, et al Testing whether barriers to a hypothetical screening test affect unrelated perceived benefits and vice versa: a randomised, experimental study. Patient Educ Couns 2017; 100(2): e1–e24.

68.

Sarma

EA.

Barriers to screening mammography. Health Psychol Rev 2015; 9(1): 42–62.

69.

Ball

Crawford

Socioeconomic status and weight change in adults: a review. Soc Sci Med 2005; 60(9): 1987–2010.

70.

Wardle

Steptoe

Socioeconomic differences in attitudes and beliefs about healthy lifestyles. J Epidemiol Community Health 2003; 57(6): 440–443.

71.

Office for National Statistics. Health state life expectancies by national deprivation deciles, England 2017 to 2019. 2021.

72.

Johnson

Deriving trends in life expectancy by the National Statistics Socioeconomic Classification using the ONS Longitudinal Study. Health Stat Q 2011; 49(49): 9–51.

73.

Benzeval

Bond

Campbell

, et al

How does money influence health?

2014.

74.

Lewer

Aldridge

Menezes

, et al Health-related quality of life and prevalence of six chronic diseases in homeless and housed people: a cross-sectional study in London and Birmingham, England. BMJ Open 2019; 9(4): e025192.

75.

Cookson

Propper

Asaria

, et al Socio-economic inequalities in health care in England. Fiscal Stud 2016; 37(3-4): 371–403.

76.

von Wagner

Baio

Raine

, et al Inequalities in participation in an organized national colorectal cancer screening programme: results from the first 2.6 million invitations in England. Int J Epidemiol 2011; 40(3): 712–718.

77.

Grosios

Gahan

Burbidge

Overview of healthcare in the UK. EPMA J 2010; 1(4): 529–534.

78.

Patel

Lawlor

Ebrahim

Socio-economic position and the use of preventive health care in older British women: a cross-sectional study using data from the British Women’s heart and Health Study cohort. Fam Pract 2007; 24(1): 7–10.

79.

Sabates

Feinstein

The role of education in the uptake of preventative health care: the case of cervical screening in britain. Soc Sci Med 2006; 62(12): 2998–3010.

80.

Morris

Baio

Kendall

, et al Socioeconomic variation in uptake of colonoscopy following a positive faecal occult blood test result: a retrospective analysis of the NHS Bowel Cancer Screening Programme. Br J Cancer 2012; 107(5): 765–771.

81.

Knight

Lindfield

The relationship between socio-economic status and access to eye health services in the UK: a systematic review. Public Health 2015; 129(2): 94–102.

82.

Carrieri

Wuebker

Assessing inequalities in preventive care use in Europe. Health Policy 2013; 113(3): 247–257.

83.

Klein

Hofreuter-Gätgens

von dem Knesebeck

. Socioeconomic status and the utilization of health services in Germany: a systematic review. In: Janssen

Swart

von Lengerke

(eds) Health Care Utilization in Germany. New York: Springer, 2014, pp.117–143.

84.

Feinstein

Sabates

Anderson

, et al What are the effects of education on health. Measuring the effects of education on health and civic engagement. In: Proceedings of the Copenhagen symposium, 2006, pp.171–354. Paris: Organisation for Economic Co-operation and Development.

85.

Kindig

Panzer

Nielsen-Bohlman

Health literacy: a prescription to end confusion. 2004.

86.

Oldach

Katz

ML.

Health literacy and cancer screening: a systematic review. Patient Educ Couns 2014; 94(2): 149–157.

87.

Kobayashi

Wardle

von Wagner

Limited health literacy is a barrier to colorectal cancer screening in England: evidence from the English Longitudinal Study of Ageing. Prev Med 2014; 61: 100–105.

88.

Bennett

Chen

Soroui

, et al The contribution of health literacy to disparities in self-rated health status and preventive health behaviors in older adults. Ann Fam Med 2009; 7: 204–211.

89.

Bolte

Tamburlini

Kohlhuber

Environmental inequalities among children in Europe–evaluation of scientific evidence and policy implications. Eur J Public Health 2010; 20(1): 14–20.

90.

Evans

Kim

Multiple risk exposure as a potential explanatory mechanism for the socioeconomic status-health gradient. Ann N Y Acad Sci 2010; 1186(1): 174–189.

91.

Evans

Kantrowitz

Socioeconomic status and health: the potential role of environmental risk exposure. Annu Rev Public Health 2002; 23(1): 303–331.

92.

Prats-Uribe

Paredes

Prieto-Alhambra

Ethnicity, comorbidity, socioeconomic status, and their associations with COVID-19 infection in England: a cohort analysis of UK Biobank data. medRxiv. 2020.

93.

Nettle

Why are there social gradients in preventative health behavior? A perspective from behavioral ecology. PLoS One 2010; 5(10): e13371.

94.

Pepper

Nettle

Out of control mortality matters: the effect of perceived uncontrollable mortality risk on a health-related decision. PeerJ 2014; 2: e459.

95.

Pepper

Nettle

Perceived extrinsic mortality risk and reported effort in looking after health. Hum Nat 2014; 25(3): 378–392.

96.

Ikeda

Iso

Toyoshima

, et al The relationships between interest for and participation in health screening and risk of mortality: the Japan Collaborative Cohort Study. Prev Med 2005; 41(3–4): 767–771.

97.

Kobayashi

von Wagner

Wardle

Perceived life expectancy is associated with colorectal cancer screening in England. Ann Behav Med 2017; 51(3): 327–336.

98.

Arndt

Routledge

Goldenberg

JL.

Predicting proximal health responses to reminders of death: the influence of coping style and health optimism. Health Psychol 2006; 21(5): 593–614.

99.

Giltay

Kamphuis

Kalmijn

, et al Dispositional optimism and the risk of cardiovascular death: the Zutphen Elderly Study. Arch Intern Med 2006; 166(4): 431–436.

100.

Vann

Tanner

Kizilova

Perceived access, fear, and preventative behavior: key relationships for positive outcomes during the covid-19 health crisis. J Consum Aff 2022; 56(1): 141–157.

101.

Gerend

Aiken

West

, et al Beyond medical risk: investigating the psychological factors underlying women’s perceptions of susceptibility to breast cancer, heart disease, and osteoporosis. Health Psychol 2004; 23(3): 247–258.

102.

Robertson

Reeve

Niedzwiedz

, et al Predictors of COVID-19 vaccine hesitancy in the UK household longitudinal study. Brain Behav Immun 2021; 94: 41–50.

103.

Woolf

McManus

Martin

, et al Ethnic differences in SARS-CoV-2 vaccine hesitancy in United Kingdom healthcare workers: results from the UK-REACH prospective nationwide cohort study. The Lancet Regional Health-Europe 2021, 9, 100180.

104.

Schmid

Rauber

Betsch

, et al Barriers of Influenza vaccination intention and behavior - A systematic review of Influenza vaccine hesitancy, 2005 - 2016. PLoS One 2017; 12(1): e0170550.

105.

Lawrenson

Bourmpaki

Bunce

, et al Trends in diabetic retinopathy screening attendance and associations with vision impairment attributable to diabetes in a large nationwide cohort. Diabetic Med 2021; 38(4): e14425.

106.

Cochrane

Gidlow

Kumar

, et al Cross-sectional review of the response and treatment uptake from the NHS Health Checks programme in Stoke on trent. J Public Health 2013; 35(1): 92–98.

107.

Nguyen

Joshi

Drew

, et al Racial and ethnic differences in COVID-19 vaccine hesitancy and uptake. medRxiv. 2021.

108.

Zintel

Flock

Arbogast

, et al Gender differences in the intention to get vaccinated against COVID-19 - a systematic review and meta-analysis. 2021.

109.

Sieverding

Zintel

Schmidt

, et al

Why do women report lower intentions to get vaccinated against COVID-19?

The mediating role of variables from the TPB and HBM. 2021.

110.

Wardle

von Wagner

Kralj-Hans

, et al Effects of evidence-based strategies to reduce the socioeconomic gradient of uptake in the English NHS Bowel Cancer Screening Programme (ASCEND): four cluster-randomised controlled trials. Lancet 2016; 387(10020): 751–759.

111.

Shickle

Farragher

TM.

Geographical inequalities in uptake of NHS-funded eye examinations: small area analysis of Leeds, UK. J Public Health 2015; 37(2): 337–345.

112.

Chiu

LF.

Inequalities of access to cancer screening: a literature review. 2003.

113.

Moser

Patnick

Beral

Inequalities in reported use of breast and cervical screening in Great Britain: analysis of cross sectional survey data. BMJ 2009; 338: b2025.

114.

Willis

Andersen

Bryant-Moore

, et al COVID-19 vaccine hesitancy: race/ethnicity, trust, and fear. Clin Transl Sci 2021; 14: 2200–2207.

115.

Hanson

Montgomery

Bakker

, et al Factors influencing mammography participation in Canada: an integrative review of the literature. Curr Oncol 2009; 16(5): 65–75.

116.

Straw

The automation of bias in medical Artificial Intelligence (AI): decoding the past to create a better future. Artif Intell Med 2020; 110: 101965.

117.

Curtis

Polygenic risk score for schizophrenia is more strongly associated with ancestry than with schizophrenia. Psychiatr Genet 2018; 28(5): 85–89.

118.

O’Brien

Lee

Carnethon

, et al Detecting dysglycemia using the 2015 United States Preventive Services Task Force screening criteria: a cohort analysis of community health center patients. PLoS Med 2016; 13(7): e1002074.

119.

Memon

Taylor

Mohebati

, et al Perceived barriers to accessing mental health services among black and minority ethnic (BME) communities: a qualitative study in southeast England. BMJ Open 2016; 6(11): e012337.

120.

Sim

Lim

, et al The perspectives of health professionals and patients on racism in healthcare: a qualitative systematic review. PLoS One 2021; 16(8): e0255936.

Predicting attitudes towards screening for neurodegenerative diseases using OCT and artificial intelligence: Findings from a literature review

Abstract

Keywords

Introduction

Predictive factors for AI-enabled screening

Properties of the disease itself and its early diagnosis

Increased understanding of neurodegenerative diseases and benefits of early diagnosis

Perceived risk of neurodegenerative disease

Social influence

Properties of the screening procedure

AI versus human agent

Experience of screening using OCT

Properties of the individual

Socioeconomic status and uncontrollable mortality risk

Age

Sex/gender

Ethnicity

Conclusion

Footnotes

Declaration of conflicting interests

Funding

Significance for public health

ORCID iDs

References