Francesc Xavier Marin-Gomez

Background: The COVID-19 pandemic disrupted primary care services worldwide, reducing quality of care. Understanding the impact of quality-improvement strategies can guide system resilience. Aim: To assess whether information provision and financial incentives to primary care professionals supported quality of care recovering to pre-pandemic levels after pandemic-related declines. Design and setting: This study involved descriptive analysis at the practice and aggregated indicator levels in Catalonia, Spain. Method: Data on 37 quality-of-care indicators from 287 primary care practices (5 250 531 adults) were analysed monthly from 2019 until 2024. Financial incentives were suspended in 2020; 13 indicators were reintegrated into pay-for-performance schemes in 2021 (early-incentivised indicators), while 24 remained out of the schemes until 2023 (late-incentivised indicators). Outcomes were annual changes in indicator results and variability, and time to recover pre-pandemic care-quality levels. A multivariate Cox model estimated the effect of incentives; Κ -means clustering identified practice recovery profiles. Results: In 2021, 11 (85%) of 13 of early-incentivised indicators started showing signs of recovery versus five (21%) of 24 late-incentivised indicators. By December 2024, recovery of pre-pandemic results occurred in 85% and 50% of early- and late-incentivised indicators, respectively, despite larger pandemic drops in early-incentivised indicators. Early incentivisation doubled the likelihood of recovery before 2023 (2.06 times), with effects varying by indicator type. However, many practices recovered late-incentivised indicators when they were informed only and before financial incentives resumed. Clustering revealed four practice profiles: fast recoverers, fast for incentivised, fast for informed, and slow recoverers. Smaller practices recovered more quickly and rural practices had milder declines. Conclusion: Financial incentives accelerated quality-of-care recovery after a system shock, particularly when tailored to indicator types and practice contexts, and complemented by timely information provision. Policymakers could use these findings to guide recovery strategies and strengthen resilience in primary care.

The prevalence of dermatological conditions in primary care, coupled with challenges such as dermatologist shortages and rising consultation costs, highlights the need for innovative solutions. Artificial intelligence (AI) holds promise for improving the diagnostic analysis of skin lesion images, potentially enhancing patient care in primary settings. This systematic review following PRISMA guidelines examined primary studies (2012–2022) assessing AI algorithms’ diagnostic accuracy for skin diseases in primary care. Studies were screened for eligibility based on their availability in the English language and exclusion criteria, with risk of bias evaluated using QUADAS-2. PubMed, Scopus, and Web of Science were searched. Fifteen studies (2019–2022), primarily from Europe and the USA, focusing on diagnostic accuracy were included. Sensitivity ranged from 58% to 96.1%, with accuracies varying from 0.41 to 0.93. AI applications encompassed triage and diagnostic support across diverse skin conditions in primary care settings, involving both patients and primary care professionals. While AI demonstrates potential for enhancing the accuracy of skin disease diagnostics in primary care, further research is imperative to address study heterogeneity and ensure algorithm reliability across diverse populations. Future investigations should prioritise robust dataset development and consider representative patient samples. Overall, AI may improve dermatological diagnosis in primary care, but careful consideration of algorithm limitations and implementation strategies is required.

Background: Artificial Intelligence (AI) has been developing for decades, but in recent years its use in the field of health care has experienced an exponential increase. Currently, there is little doubt that these tools have transformed clinical practice. Therefore, it is important to know how the population perceives its implementation to be able to propose strategies for acceptance and implementation and to improve or prevent problems arising from future applications. Objective: This study aims to describe the population’s perception and knowledge of the use of AI as a health support tool and its application to radiology through a validated questionnaire, in order to develop strategies aimed at increasing acceptance of AI use, reducing possible resistance to change and identifying possible sociodemographic factors related to perception and knowledge. Methods: A cross-sectional observational study was conducted using an anonymous and voluntarily validated questionnaire aimed at the entire population of Catalonia aged 18 years or older. The survey addresses 4 dimensions defined to describe users’ perception of the use of AI in radiology, (1) “distrust and accountability,” (2) “personal interaction,” (3) “efficiency,” and (4) “being informed,” all with questions in a Likert scale format. Results closer to 5 refer to a negative perception of the use of AI, while results closer to 1 express a positive perception. Univariate and bivariate analyses were performed to assess possible associations between the 4 dimensions and sociodemographic characteristics. Results: A total of 379 users responded to the survey, with an average age of 43.9 (SD 17.52) years and 59.8% (n=226) of them identified as female. In addition, 89.8% (n=335) of respondents indicated that they understood the concept of AI. Of the 4 dimensions analyzed, “distrust and accountability” obtained a mean score of 3.37 (SD 0.53), “personal interaction” obtained a mean score of 4.37 (SD 0.60), “efficiency” obtained a mean score of 3.06 (SD 0.73) and “being informed” obtained a mean score of 3.67 (SD 0.57). In relation to the “distrust and accountability” dimension, women, people older than 65 years, the group with university studies, and the population that indicated not understanding the AI concept had significantly more distrust in the use of AI. On the dimension of “being informed,” it was observed that the group with university studies rated access to information more positively and those who indicated not understanding the concept of AI rated it more negatively. Conclusions: The majority of the sample investigated reported being familiar with the concept of AI, with varying degrees of acceptance of its implementation in radiology. It is clear that the most conflictive dimension is “personal interaction,” whereas “efficiency” is where there is the greatest acceptance, being the dimension in which there are the best expectations for the implementation of AI in radiology.

BACKGROUND Artificial Intelligence (AI) has been developing for decades, but in recent years its use in the field of health care has experienced an exponential increase. Currently, there is little doubt that these tools have transformed clinical practice. Therefore, it is important to know how the population perceives its implementation to be able to propose strategies for acceptance and implementation and to improve or prevent problems arising from future applications. OBJECTIVE This study aims to describe the population’s perception and knowledge of the use of AI as a health support tool and its application to radiology through a validated questionnaire, in order to develop strategies aimed at increasing acceptance of AI use, reducing possible resistance to change and identifying possible sociodemographic factors related to perception and knowledge. METHODS A cross-sectional observational study was conducted using an anonymous and voluntarily validated questionnaire aimed at the entire population of Catalonia aged 18 years or older. The survey addresses 4 dimensions defined to describe users’ perception of the use of AI in radiology, (1) “distrust and accountability,” (2) “personal interaction,” (3) “efficiency,” and (4) “being informed,” all with questions in a Likert scale format. Results closer to 5 refer to a negative perception of the use of AI, while results closer to 1 express a positive perception. Univariate and bivariate analyses were performed to assess possible associations between the 4 dimensions and sociodemographic characteristics. RESULTS A total of 379 users responded to the survey, with an average age of 43.9 (SD 17.52) years and 59.8% (n=226) of them identified as female. In addition, 89.8% (n=335) of respondents indicated that they understood the concept of AI. Of the 4 dimensions analyzed, “distrust and accountability” obtained a mean score of 3.37 (SD 0.53), “personal interaction” obtained a mean score of 4.37 (SD 0.60), “efficiency” obtained a mean score of 3.06 (SD 0.73) and “being informed” obtained a mean score of 3.67 (SD 0.57). In relation to the “distrust and accountability” dimension, women, people older than 65 years, the group with university studies, and the population that indicated not understanding the AI concept had significantly more distrust in the use of AI. On the dimension of “being informed,” it was observed that the group with university studies rated access to information more positively and those who indicated not understanding the concept of AI rated it more negatively. CONCLUSIONS The majority of the sample investigated reported being familiar with the concept of AI, with varying degrees of acceptance of its implementation in radiology. It is clear that the most conflictive dimension is “personal interaction,” whereas “efficiency” is where there is the greatest acceptance, being the dimension in which there are the best expectations for the implementation of AI in radiology.

BACKGROUND The use of Artificial intelligence (AI) algorithms for detecting different ophthalmic diseases, especially diabetic retinopathy (DR), has become increasingly popular. OBJECTIVE To evaluate the screening performance of different AI algorithms based on convolutional neural networks (CNNs) in a real-world scenario METHODS Observational and cross-sectional study conducted on patients aged ≥18 years with type-2 diabetes mellitus, who had undergone fundus examination for DR screening using a teleophthalmology program between May and August 2021 in four Catalan primary care centers. Color fundus photos were acquired, using a non-mydriatic camera, and transmitted to the reading center of the ophthalmological clinic. We used the UPRETINA diagnostic system, which consists of 8 AI algorithms based on CNN's. Sensitivity and specificity; inter- and intra-observer agreement; and area under the receiver operating characteristic curve (AUROC) were calculated. RESULTS A total of 1652 eyes from 871 patients [1652 images] were analyzed. The AI algorithms had a sensitivity/specificity of 86.8%/95.6% for detecting DR; 94.9%/94.3% for detecting age-related macular degeneration (AMD); 82.7%/92.4% for detecting glaucomatous optic neuropathy (GON); 87.0%/87.5% for detecting epiretinal membrane; and 89.7%/98.0% for detecting nevus. Additionally, the sensitivity/specificity for correctly classify images as right eye/left eye and to correctly classify images gradeability (medium or high quality) were 100% /100% and 92.9%/90.5%, respectively. The AUROC of the AI algorithms ranged between 0.9777 (AMD) and 0.9122 (GON). CONCLUSIONS UPRETINA system algorithms were capable to automatically and accurately classify the screening retinographies, which obviously have a significant impact on reducing workload. In addition, it has been shown to be able to distinguish pathological images from non-pathological ones, leading to a scenario of more efficient optimization of resources, in which retina specialists will be focused on assessing only pathological images. This would also reduce primary-care burden. The incorporation of a greater spectrum of data, always maintaining and guaranteeing patient confidentiality, will help to establish a patient-tailored health care. CLINICALTRIAL ClinicalTRials. Gov NCT04132401. INTERNATIONAL REGISTERED REPORT RR2-10.2196/12539

BACKGROUND Dermatological conditions are a relevant health problem. Machine learning (ML) models are increasingly being applied to dermatology as a diagnostic decision support tool using image analysis, especially for skin cancer detection and disease classification. OBJECTIVE The objective of this study was to perform a prospective validation of an image analysis ML model, which is capable of screening 44 skin diseases, comparing its diagnostic accuracy with that of General Practitioners (GPs) and teledermatology (TD) dermatologists in a real-life setting. METHODS Prospective, diagnostic accuracy study including 100 consecutive patients with a skin problem who visited a participating GP in central Catalonia, Spain, between June 2021 and October 2021. The skin issue was first assessed by the GPs. Then an anonymised skin disease picture was taken and uploaded to the ML application, which returned a list with the Top-5 possible diagnosis in order of probability. The same image was then sent to a dermatologist via TD for diagnosis, as per clinical practice. The GPs Top-3, ML model’s Top-5 and dermatologist’s Top-3 assessments were compared to calculate the accuracy, sensitivity, specificity and diagnostic accuracy of the ML models. RESULTS The overall Top-1 accuracy of the ML model (39%) was lower than that of GPs (64%) and dermatologists (72%). When the analysis was limited to the diagnoses on which the algorithm had been explicitly trained (n=82), the balanced Top-1 accuracy of the ML model increased (48%) and in the Top-3 (75%) was comparable to the GPs Top-3 accuracy (76%). The Top-5 accuracy of the ML model (89%) was comparable to the dermatologist Top-3 accuracy (90%). For the different diseases, the sensitivity of the model (Top-3 87% and Top-5 84%) is higher than that of the clinicians (Top-3 GPs 76% and Top-3 dermatologists 84%) only in the benign tumour pathology group, being on the other hand the most prevalent category (n=53). About the satisfaction of professionals, 92% of the GPs considered it as a usefuldiagnostic support tool (DST) for the differential diagnosis and in 60% of the cases as an aid in the final diagnosis of the skin lesion. CONCLUSIONS The overall diagnostic accuracy of the model in this study under real conditions is lower than that of both GPs and dermatologists, a fact that is consistent with the few existing prospective studies under real conditions. These results highlight the potential of the ML models to assist GPs as a DST for skin conditions especially in the differential diagnosis. However, external testing in real conditions is essential for data validation and regulating these AI diagnostic models, in order to deploy ML models in a Primary Care setting. CLINICALTRIAL The clinical trial has been approved by the IDIAP Jordi Gol i Gurina ethics committee with code 20-159P. Clinicaltrials.gov NCT04562168; https://clinicaltrials.gov/ct2/show/NCT04562168 (Archived by WebCite at http://www.webcitation.org/6V0tto0lZ). JMIR Res Protoc 2022;11(8):e37531 doi:10.2196/37531

Background: Tobacco consumption during pregnancy is one of the most modifiable causes of morbidity and mortality for both pregnant smokers and their foetus. Even though pregnant smokers are conscious about the negative effects of tobacco consumption, they also had barriers for smoking cessation and most of them continue smoking, being a major public health problem. The aim of this study is to determine the effectiveness of an application (App) for mobile devices, designed with a gamification strategy, in order to help pregnant smokers to quit smoking during pregnancy and in the long term. Methods: This study is a multicentre randomized community intervention trial. It will recruit pregnant smokers (200 participants/group), aged more than 18 years, with sporadically or daily smoking habit in the last 30 days and who follow-up their pregnancy in the Sexual and Reproductive Health Care Services of the Camp de Tarragona and Central Catalonia Primary Care Departments. All the participants will have the usual clinical practice intervention for smoking cessation, whereas the intervention group will also have access to the App. The outcome measure will be prolonged abstinence at 12 months after the intervention, as confirmed by expired-carbon monoxide and urinary cotinine tests. Results will be analysed based on intention to treat. Prolonged abstinence rates will be compared, and the determining factors will be evaluated using multivariate statistical analysis. Discussion: The results of this study will offer evidence about the effectiveness of an intervention using a mobile App in smoking cessation for pregnant smokers, to decrease comorbidity associated with long-term smoking. If this technology is proven effective, it could be readily incorporated into primary care intervention for all pregnant smokers. Trial registration: Clinicaltrials.gov ID NCT05222958. Trial registered 3 February 2022.

Background: Dermatological conditions are a relevant health problem. Each person has an average of 1.6 skin diseases per year, and consultations for skin pathology represent 20% of the total annual visits to primary care and around 35% are referred to a dermatology specialist. Machine learning (ML) models can be a good tool to help primary care professionals, as it can analyze and optimize complex sets of data. In addition, ML models are increasingly being applied to dermatology as a diagnostic decision support tool using image analysis, especially for skin cancer detection and classification. Objective: This study aims to perform a prospective validation of an image analysis ML model as a diagnostic decision support tool for the diagnosis of dermatological conditions. Methods: In this prospective study, 100 consecutive patients who visit a participant general practitioner (GP) with a skin problem in central Catalonia were recruited. Data collection was planned to last 7 months. Anonymized pictures of skin diseases were taken and introduced to the ML model interface (capable of screening for 44 different skin diseases), which returned the top 5 diagnoses by probability. The same image was also sent as a teledermatology consultation following the current stablished workflow. The GP, ML model, and dermatologist’s assessments will be compared to calculate the precision, sensitivity, specificity, and accuracy of the ML model. The results will be represented globally and individually for each skin disease class using a confusion matrix and one-versus-all methodology. The time taken to make the diagnosis will also be taken into consideration. Results: Patient recruitment began in June 2021 and lasted for 5 months. Currently, all patients have been recruited and the images have been shown to the GPs and dermatologists. The analysis of the results has already started. Conclusions: This study will provide information about ML models’ effectiveness and limitations. External testing is essential for regulating these diagnostic systems to deploy ML models in a primary care practice setting.

Nursing homes have accounted for a significant part of SARS-CoV-2 mortality, causing great social alarm. Using data collected from electronic medical records of 1,319,839 institutionalised and non-institutionalised persons ≥ 65 years, the present study investigated the epidemiology and differential characteristics between these two population groups. Our results showed that the form of presentation of the epidemic outbreak, as well as some risk factors, are different among the elderly institutionalised population with respect to those who are not. In addition to a twenty-fold increase in the rate of adjusted mortality among institutionalised individuals, the peak incidence was delayed by approximately three weeks. Having dementia was shown to be a risk factor for death, and, unlike the non-institutionalised group, neither obesity nor age were shown to be significantly associated with the risk of death among the institutionalised. These differential characteristics should be able to guide the actions to be taken by the health administration in the event of a similar infectious situation among institutionalised elderly people.

Background: The COVID-19 pandemic has turned the care model of health systems around the world upside down, causing the abrupt cancellation of face-to-face visits and redirection of the model toward telemedicine. Digital transformation boosts information systems—the more robust they are, the easier it is to monitor the health care system in a highly complex state and allow for more agile and reliable analysis. Objective: The purpose of this study was to analyze diagnoses from primary care visits and distinguish between those that had higher and lower variations, relative to the 2019 and 2020 periods (roughly before and during COVID-19), to identify clinical profiles that may have been most impaired from the least-used diagnostic codes for visits during the pandemic. Methods: We used a database from the Primary Care Services Information Technologies Information System of Catalonia. We analyzed the register of visits (n=2,824,185) and their International Classification of Diseases (ICD-10) diagnostic codes (n=3,921,974; mean 1.38 per visit), as approximations of the reasons for consultations, at 3 different grouping levels. The data were represented by a term frequency matrix and analyzed recursively in different partitions aggregated according to date. Results: The increase in non–face-to-face visits (+267%) did not counterbalance the decrease in face-to-face visits (–47%), with an overall reduction in the total number of visits of 1.36%, despite the notable increase in nursing visits (10.54%). The largest increases in 2020 were visits with diagnoses related to COVID-19 (ICD-10 codes Z20-Z29: 2.540%), along with codes related to economic and housing problems (ICD-10 codes Z55-Z65: 44.40%). Visits with most of the other diagnostic codes decreased in 2020 relative to those in 2019. The largest reductions were chronic pathologies such as arterial hypertension (ICD-10 codes I10-I16: –32.73%) or diabetes (ICD-10 codes E08-E13: –21.13%), but also obesity (E65-E68: –48.58%) and bodily injuries (ICD-10 code T14: –33.70%). Visits with mental health–related diagnostic codes decreased, but the decrease was less than the average decrease. There was a decrease in consultations—for children, adolescents, and adults—for respiratory infections (ICD-10 codes J00-J06: –40.96%). The results show large year-on-year variations (in absolute terms, an average of 12%), which is representative of the strong shock to the health system. Conclusions: The disruption in the primary care model in Catalonia has led to an explosive increase in the number of non–face-to-face visits. There has been a reduction in the number of visits for diagnoses related to chronic pathologies, respiratory infections, obesity, and bodily injuries. Instead, visits for diagnoses related to socioeconomic and housing problems have increased, which emphasizes the importance of social determinants of health in the context of this pandemic. Big data analytics with routine care data yield findings that are consistent with those derived from intuition in everyday clinical practice and can help inform decision making by health planners in order to use the next few years to focus on the least-treated diseases during the COVID-19 pandemic.