Modelling presentation delay in stroke—what are we learning?

This invited commentary refers to ‘Pre-hospital delay intention and its associated factors in the high-risk population of stroke: A latent profile analysis’ by M. Chen et al., https://doi.org/10.1093/eurjcn/zvae136. The prevalence of stroke is predicted to rise to 6.4% of adults in the USA by 2050.1 This increase occurs in tandem with a surge in the prevalence of total cardiovascular disease (CVD), with 15% of US adults (45 million people) predicted to develop some form of CVD by 2050.1 These figures are staggering and hammer home the need for strategic public health interventions that prevent both diseases. An important tenet of public health interventions is national educational campaigns.1 In this respect, public health leaders working in stroke have had considerable success with campaigns like the FAST (Face, Arm, Speech, Time) mnemonic tool that aimed to educate about stroke symptoms.2 Successive studies in the USA, Canada, and the UK, where the FAST initiative has been implemented, have demonstrated the mnemonic tool leads to overall improved knowledge of stroke, earlier recognition of symptoms, and even increased thrombolysis rates.2,3 Despite the aforementioned success, delayed presentation remains a major problem in stroke.4,5 Therefore, there is a need to assess, evolve, and renew efforts in order to continue to drive improvements. Updated versions of FAST like BE-FAST, which added the ‘B’ for balance and ‘E’ for eyes,6 have been proposed as a means of increasing the sensitivity of the FAST acronym. Interestingly, subsequent research has demonstrated that whilst ‘BE-FAST’ may have greater sensitivity at detecting stroke in a retrospective case review, it is not superior to the FAST mnemonic at prospectively identifying stroke, and potentially results in lower retention of the symptom knowledge the acronym aims to instil.2 The overall message from epidemiological studies,5 evaluations of public health endeavours,2 and reviews of factors that result in delayed thrombolysis4 is that more work is needed to improve the public’s ability to quickly and accurately identify and respond to the symptoms of stroke. In this issue, Chen et al. have stepped up to this research challenge. Using a large sample of predominantly male high-risk stroke individuals in China (n = 457), the authors have performed a latent profile analysis to define distinct stroke ‘help-seeking’ phenotypes that are based on responses to the Stroke Pre-hospital Delay Behavior Intention.7 The latter is a valid measure of the likeliness of pre-hospital delay in high-risk stroke patients and their family members.8 Latent profile or class analyses are useful as they allow researchers to identify ‘homogenous’ groups amongst what appears to be significant heterogeneity.9 The mixture of probabilistic calculations and mathematical testing of ‘fit’ can allow researchers to observe the hitherto unobserved. In this study, the four-class model was selected as the optimal model, and the profile categories were labelled: Class 1: high warning signs with low delay intention (26%); Class 2: low warning signs with low delay intention (18%); Class 3: moderate level of delay intention (37%); Class 4: high level of delay intention (19%). The authors then employed a multinomial logistic regression analysis to interrogate which variables were associated with membership of each class (in reference to Class 4). The predictor variables (n = 19) were a priori derived from a comprehensive literature review. Higher levels of education, closer proximity to medical services, and higher scores on the stroke knowledge questionnaire and health belief questionnaire were associated with a higher probability of membership to Class 1 over Class 4. Higher income and higher stroke knowledge scores were associated with a higher probability of membership of Class 2 compared with Class 4. Lower age categories were more likely to be associated with a higher probability of membership to Class 3 compared with Class 4. The authors concluded that this information may be help healthcare workers identify those at greater risk of delayed presentation and aid in the tailoring or personalization of preventative interventions.7 The study by Chen et al.7 adds to the large body of literature of the influence of disparities in delay times to hospital presentation with stroke.4 Given the potentially critical importance of timely hospital treatment such as thrombolytic therapy,2,3 these disparities will lead to significant health inequalities. In recognition of this, Chen et al. conclude that their study may help healthcare professionals develop targeted interventions to promote health behaivour for each subgroup. Addressing health inequalities is highly complex, and any future intervention is likely to require a systems approach that enables appreciation of the complex interactions between the individual and environmental determinants of health and behaviour.10 The work of Chen et al. is of great value; using a latent profile analysis, they have identified potentially important profiles that may not have been obvious through conventional analyses. Further, this research has highlighted some important factors that might be relevant to advancing our understanding of delays in seeking treatment in stroke. Future programmes of work seeking to address these identified inequalities may utilize these findings to develop a deeper theoretical understanding of the complexity of issues contributing to low levels of intent to seek medical attention with symptoms of stroke. In addition, they may also be valuable to researchers who are venturing to design solutions using system approaches. Stroke can have a devastating impact on the individuals and their family, and early treatment is essential to optimize the chance of recovery and reduced the likelihood of permanent disability. Any research that highlights potential factors that contribute to delayed help seeking, such as presented by Chen et al., is therefore extremely valuable to advancing the field. Faye Forsyth (Conceptualization, Validation, Writing—original draft, Writing—review and editing [equal]) and Peter Hartley (Conceptualization, Validation, Writing—original draft, Writing—review and editing [equal]). No funding was involved in this publication. Data sharing is not applicable as no new data were used in this invited commentary.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.