Patrícia Pita Ferreira

Dr. Patrícia Pita Ferreira

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
22 November 2023 in JMIR AI
Verified
Journal Article
Open access
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
DSDiogo Godinho Simões
Diogo Godinho Simões
CCConstança Pinto de Carvalho
Constança Pinto de Carvalho
FDFrancisco Duarte
Francisco Duarte
EFEugénia Fernandes
Eugénia Fernandes
APAndré Peralta-Santos
André Peralta-Santos
Background: In 2021, the European Union reported >270,000 excess deaths, including >16,000 in Portugal. The Portuguese Directorate-General of Health developed a deep neural network, AUTOCOD, which determines the primary causes of death by analyzing the free text of physicians’ death certificates (DCs). Although AUTOCOD’s performance has been established, it remains unclear whether its performance remains consistent over time, particularly during periods of excess mortality. Objective: This study aims to assess the sensitivity and other performance metrics of AUTOCOD in classifying underlying causes of death compared with manual coding to identify specific causes of death during periods of excess mortality. Methods: We included all DCs between 2016 and 2019. AUTOCOD’s performance was evaluated by calculating various performance metrics, such as sensitivity, specificity, positive predictive value (PPV), and F1-score, using a confusion matrix. This compared International Statistical Classification of Diseases and Health-Related Problems, 10th Revision (ICD-10), classifications of DCs by AUTOCOD with those by human coders at the Directorate-General of Health (gold standard). Subsequently, we compared periods without excess mortality with periods of excess, severe, and extreme excess mortality. We defined excess mortality as 2 consecutive days with a Z score above the 95% baseline limit, severe excess mortality as 2 consecutive days with a Z score >4 SDs, and extreme excess mortality as 2 consecutive days with a Z score >6 SDs. Finally, we repeated the analyses for the 3 most common ICD-10 chapters focusing on block-level classification. Results: We analyzed a large data set comprising 330,098 DCs classified by both human coders and AUTOCOD. AUTOCOD demonstrated high sensitivity (≥0.75) for 10 ICD-10 chapters examined, with values surpassing 0.90 for the more prevalent chapters (chapter II—“Neoplasms,” chapter IX—“Diseases of the circulatory system,” and chapter X—“Diseases of the respiratory system”), accounting for 67.69% (223,459/330,098) of all human-coded causes of death. No substantial differences were observed in these high-sensitivity values when comparing periods without excess mortality with periods of excess, severe, and extreme excess mortality. The same holds for specificity, which exceeded 0.96 for all chapters examined, and for PPV, which surpassed 0.75 in 9 chapters, including the more prevalent ones. When considering block classification within the 3 most common ICD-10 chapters, AUTOCOD maintained a high performance, demonstrating high sensitivity (≥0.75) for 13 ICD-10 blocks, high PPV for 9 blocks, and specificity of >0.98 in all blocks, with no significant differences between periods without excess mortality and those with excess mortality. Conclusions: Our findings indicate that, during periods of excess and extreme excess mortality, AUTOCOD’s performance remains unaffected by potential text quality degradation because of pressure on health services. Consequently, AUTOCOD can be dependably used for real-time cause-specific mortality surveillance even in extreme excess mortality situations.

24 January 2024 in Vaccines
Verified
Journal Article
Open access
PDF file
CCConstança Pinto de Carvalho
Constança Pinto de Carvalho
MRManuel Ribeiro
Manuel Ribeiro
DSDiogo Godinho Simões
Diogo Godinho Simões
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
LALeonardo Azevedo
Leonardo Azevedo
APAndré Peralta-Santos
André Peralta-Santos
Vaccine hesitancy tends to exhibit geographical patterns and is often associated with social deprivation and migrant status. We aimed to estimate COVID-19 vaccination hesitancy in a high-vaccination-acceptance country, Portugal, and determine its association with sociodemographic risk factors. We used the Registry of National Health System Users to determine the eligible population and the Vaccination Registry to determine individuals without COVID-19 vaccine doses. Individuals older than five with no COVID-19 vaccine dose administered by 31 March 2022 were considered hesitant. We calculated hesitancy rates by municipality, gender, and age group for all municipalities in mainland Portugal. We used the spatial statistical scan method to identify spatial clusters and the Besag, Yorke, and Mollié (BYM) model to estimate the effect of age, gender, social deprivation, and migrant proportion across all mainland municipalities. The eligible population was 9,852,283, with 1,212,565 (12%) COVID-19 vaccine-hesitant individuals. We found high-hesitancy spatial clusters in the Lisbon metropolitan area and the country’s southwest. Our model showed that municipalities with higher proportions of migrants are associated with an increased relative risk (RR) of vaccine hesitancy (RR = 8.0; CI 95% 4.6; 14.0). Social deprivation and gender were not associated with vaccine hesitancy rates. We found COVID-19 vaccine hesitancy has a heterogeneous distribution across Portugal and has a strong association with the proportion of migrants per municipality.

6 November 2025 in medRxiv
Verified
Preprint
Open access
PDF file
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
SLSara Soriano Longarón
Sara Soriano Longarón
WBWiam Bouisaghouane
Wiam Bouisaghouane
JGJetse Goris
Jetse Goris
AHAnne H. Hoekman
Anne H. Hoekman
MPMirjam Plantinga
Mirjam Plantinga
Artificial intelligence (AI) in healthcare holds transformative potential but risks exacerbating existing health disparities if inclusivity is not explicitly accounted for. This study addresses the disconnected discussions on inclusive medical AI by developing a comprehensive framework, PREFER-IT. This framework is based on the outcomes of a five-day transdisciplinary co-creation workshop that involved 37 experts from diverse backgrounds, including healthcare, ethics, law, social sciences, AI, and patient advocacy. For this workshop, we used design thinking and participatory methodologies to develop a framework for realising inclusive medical AI. We identified three key challenges for realising inclusive medical AI: integrating the lived experiences and stakeholder voices across the AI lifecycle, designing data collection practices that promote fairness and prevent inequalities, and fostering regulatory frameworks to uphold human rights and promote inclusivity. The analysis of participants’ perspectives informed the development of eight key thematic clusters of PREFER-IT: Participatory and co-design approaches (P), Representative and diverse data (R), Education and digital literacy (E), Fairness (F), Ethical and legal accountability (E), Real-world validation and feedback (R), Inclusive communication (I), and Technical interoperability (T). These elements were mapped across structural layers of AI (humans, data, system, process, and governance) and the AI lifecycle to guide inclusive design, development, validation, implementation, monitoring, and governance. This framework fosters stakeholder engagement and systemic change, positioning inclusion as a guiding principle in practice. PREFER-IT offers a practical and conceptual contribution for how to include ethical, legal and societal aspects when aiming to foster responsible and inclusive AI in healthcare. Author Summary: Artificial intelligence (AI) is being used more and more in healthcare to improve diagnosis, treatment, and personalised care. However, if not designed carefully, these technologies can unintentionally increase existing inequalities and exclude certain groups from their benefits. In our study, we brought together experts from healthcare, ethics, law, social sciences, and patient advocacy to explore how AI in medicine can be made more inclusive. Over five days, we worked together to identify key issues and come up with practical solutions. We focused on three main areas: 1) Ensuring diverse voices are heard during the development of AI tools; 2) Making data collection fair and representative; and 3) Creating regulations that protect human rights. From the discussions of the workshop, we created the PREFER-IT framework, which outlines eight key principles for inclusive AI: P articipatory and co-design approaches R epresentative and diverse data E ducation and digital literacy F airness E thical and legal accountability R eal-world validation and feedback I nclusive communication T echnical interoperability This framework helps guide developers, policymakers, and healthcare professionals in creating AI systems that are not only effective but also fair and respectful of all users. Our work emphasises the importance of involving patients and communities in shaping the future of AI.

20 October 2021 in European Journal of Public Health
Verified
Journal Article
Open access
PDF file
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
RMR Martins
R Martins
DTDp Tavares
Dp Tavares
JVJ Valente
J Valente
Issue/problem: At the time of arrival of the COVID-19 pandemic to Portugal, the Public Health Department of Lisbon and Tagus Valley Regional Health Administration was confronted with its obsolete information system, which could not provide the information that Public Health Authorities needed to bring the pandemic under control. In order to tackle this issue, our goal was to aggregate all the relevant epidemiological data for the region in an accessible and user-friendly dashboard. Methods: Starting from March 2020 we created a dashboard, which compiled and analyzed COVID-19 data, such as: confirmed and active cases, deaths, hospitalizations (including ICU beds), high/low risk contacts, mandatory confinements, tests, and clusters. This dashboard had anonymized data and was updated daily as well as constantly upgraded by taking into consideration new incoming data and feedback provided by users, serving its purpose as an advisory tool for policy. Results: The dashboard was available to Local, Regional, and National Public Health Authorities and made possible to monitor trends and predict epidemiological changes that, prior to its existence, were dispersed in several datasets. Most importantly, it allowed for a better allocation of human resources for contact tracing and case management activities, to easily identify basic social needs for the most vulnerable citizens. It also provided for policy measures adjusted to the smallest administrative division in Portugal, the parish, that ultimately allowed for a better epidemiological control in Lisbon Metropolitan Area in 2020. Lessons: The regional dashboard is far from perfect, but it highlights the constant need for robust information and epidemiological surveillance systems, serving both National, Regional and Local Public Health Authorities. It appeared out of necessity during a pandemic and it proved once more the importance of data analytics for guiding public health action based on evidence. Key messages: Data analytics can serve both as an advisory tool for policy as well as to guide public health interventions. The COVID-19 Dashboard for the Lisbon Region aggregates relevant epidemiological data that helped Public Health authorities better understand and control the Pandemic in the region.

11 July 2022 in JMIR Preprints
Verified
Preprint
Open access
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
DSDiogo Godinho Simões
Diogo Godinho Simões
CCConstança Pinto de Carvalho
Constança Pinto de Carvalho
FDFrancisco Duarte
Francisco Duarte
EFEugénia Fernandes
Eugénia Fernandes
APAndré Peralta-Santos
André Peralta-Santos
BACKGROUND In 2021, the European Union reported >270,000 excess deaths, including >16,000 in Portugal. The Portuguese Directorate-General of Health developed a deep neural network, AUTOCOD, which determines the primary causes of death by analyzing the free text of physicians’ death certificates (DCs). Although AUTOCOD’s performance has been established, it remains unclear whether its performance remains consistent over time, particularly during periods of excess mortality. OBJECTIVE This study aims to assess the sensitivity and other performance metrics of AUTOCOD in classifying underlying causes of death compared with manual coding to identify specific causes of death during periods of excess mortality. METHODS We included all DCs between 2016 and 2019. AUTOCOD’s performance was evaluated by calculating various performance metrics, such as sensitivity, specificity, positive predictive value (PPV), and F1-score, using a confusion matrix. This compared International Statistical Classification of Diseases and Health-Related Problems, 10th Revision (ICD-10), classifications of DCs by AUTOCOD with those by human coders at the Directorate-General of Health (gold standard). Subsequently, we compared periods without excess mortality with periods of excess, severe, and extreme excess mortality. We defined excess mortality as 2 consecutive days with a Z score above the 95% baseline limit, severe excess mortality as 2 consecutive days with a Z score >4 SDs, and extreme excess mortality as 2 consecutive days with a Z score >6 SDs. Finally, we repeated the analyses for the 3 most common ICD-10 chapters focusing on block-level classification. RESULTS We analyzed a large data set comprising 330,098 DCs classified by both human coders and AUTOCOD. AUTOCOD demonstrated high sensitivity (≥0.75) for 10 ICD-10 chapters examined, with values surpassing 0.90 for the more prevalent chapters (chapter II—“Neoplasms,” chapter IX—“Diseases of the circulatory system,” and chapter X—“Diseases of the respiratory system”), accounting for 67.69% (223,459/330,098) of all human-coded causes of death. No substantial differences were observed in these high-sensitivity values when comparing periods without excess mortality with periods of excess, severe, and extreme excess mortality. The same holds for specificity, which exceeded 0.96 for all chapters examined, and for PPV, which surpassed 0.75 in 9 chapters, including the more prevalent ones. When considering block classification within the 3 most common ICD-10 chapters, AUTOCOD maintained a high performance, demonstrating high sensitivity (≥0.75) for 13 ICD-10 blocks, high PPV for 9 blocks, and specificity of >0.98 in all blocks, with no significant differences between periods without excess mortality and those with excess mortality. CONCLUSIONS Our findings indicate that, during periods of excess and extreme excess mortality, AUTOCOD’s performance remains unaffected by potential text quality degradation because of pressure on health services. Consequently, AUTOCOD can be dependably used for real-time cause-specific mortality surveillance even in extreme excess mortality situations.

1 October 2023 in European Journal of Public Health
Verified
Journal Article
Open access
PDF file
Patrícia Pita FerreiraPatrícia Pita Ferreira
Patrícia Pita Ferreira
RMR Martins
R Martins
AFA R Filipe
A R Filipe
CLC Lopes
C Lopes
CCC Costa
C Costa
APA C Pecante
A C Pecante
Issue: The World Health Organization considers viral hepatitis a public health problem and aims to eliminate it as a threat by 2030, reducing incidence by 90% and mortality by 65%. To achieve these goals, the National Program for Viral Hepatitis recommends the screening of viral hepatitis at least once in a lifetime. However, this is not a widespread practice. Description of the problem: Micro-elimination is an effective, efficient, and cost-effective intervention in addressing chronic hepatitis in specific geographic areas. The ‘Hep-ØN’ project aims to reduce the incidence, mortality, and morbidity of viral hepatitis in the area of influence of Oeste Norte. The project has three strategic axes: screening, patient follow-up, and liver disease literacy, with the support of several partners. Results: The goal is to implement viral hepatitis screening at least once in a lifetime for all adult users, improve referral networks, ensure treatment and post-treatment surveillance of viral hepatitis, and empower healthcare professionals. The project aims to cover the 180,000 users registered in Oeste Norte (municipalities of Alcobaça, Bombarral, Caldas da Rainha, Óbidos, Peniche, Nazaré), of which more than 150,000 are over 18 years old. The pilot project is taking place in Peniche until September 2023, and covers more than 27,000 users, of which more than 22,000 are adults. The project's strategies can be applied to other areas, allowing for the rapid identification and treatment of viral hepatitis. Lessons: The success of the pilot project in Peniche will demonstrate the importance of integrated and localized interventions in preventing and eliminating chronic viral hepatitis, first in ACES ON and then in other regions of the country. Replication should be adapted to local characteristics, leveraging synergies with local and national partners. Integrated care models are an effective strategy to address the challenges of preventing, treating, and controlling chronic viral hepatitis. Key messages: • ‘Hep-ØN’ project aims to screen all adults, treat viral hepatitis, and empower healthcare professionals in Oeste Norte to eliminate chronic hepatitis. • Success of the pilot project in Peniche will show localized interventions’ importance in identifying and treating viral hepatitis rapidly and how integrated care models are an effective strategy.