Dr. Vered Bar, Head of the Women’s Mental Health at Sheba Medical Center, presented at the 17th national meeting of the Israeli Psychiatric Association a joint work with KI’s researchers titled “Estimation of Postpartum Depression Risk from Electronic Health Records Using Machine Learning”Read Less
KI senior researcher Guy Amit will present our work on predicting post-partum depression. The first ever virtual congress of the European Psychiatric Association was attended by more than 2,900 people. KI’s research with collaborators from Weil-Cornell Medicine and Sheba medical center was presented in the oral session titled ‘Depressive Disorders’.Read Less
KI Senior researcher Chen Yanover participated in the virtual OHDSI COVID-19 Study-A-Thon. Chen participated in a COVID-19 global, virtual study-a-thon of the Observational Health Data Sciences and Informatics (OHDSI) community. More than 330 people from over 30 countries collaborated in the event, aiming to make an impact on the current global pandemic by bringing out value from COVID-19 health data analytics.Read Less
This study came to address the uncertainty regarding the long-term effects of the pandemic and the fear that a large proportion of infected individuals will have long-lasting symptoms and emergence of new morbidities.
In a collaborative effort with our incredible partners from Maccabi Health care Services, we analyzed electronic health records of almost 300K infected individuals matched carefully to controls and measured the health outcomes that lingered a year from a mild COVID-19.
We found that only a few symptoms persisted a year from a mild SARS-CoV-2 infection and their risk decreased with time from infection. Children had fewer outcomes, which mostly resolved in the second half of the year, sex had a minor effect on risk of outcomes, and findings remained consistent across SARS-CoV-2 variants. Vaccinated patients with breakthrough covid-19 infection had a lower risk for shortness of breath and similar risk for other outcomes compared with unvaccinated patients.
Our study could contribute to other global efforts in assessing the magnitude of long COVID to allow budget allocation to patients who need them and enable physicians to identify and better treat lingering outcomes following SARS-CoV-2 infection. This will also provide invaluable knowledge to the public of what to expect following COVID-19 illness and reduce the uncertainty around it.
This unique scale has been developed by KI Research Institute, The Israeli Ministry of Health and the Child Development Institute in the Negev. It is based on 3.5 million developmental examinations performed on Israeli children and provides new guidelines for optimal acquisition of developmental milestones. Just a few months after our initial publication (https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2790042), it is now widely spread and used by a variety of child development professionals. The scale is the new Israeli standard for nurses in our national child development centers , early childhood education teams, pediatricians and parents, covering all developmental steps in the first 5 years of life and promoting early interventions. As we strive to create impact and advance our findings and the tools we build towards implementation, this is a major developmental milestone achieved by our young institute.Read Less
Creating the first evidence-based developmental scale of Israeli children has been one of our most important projects for the past two years. Today, the manuscript describing this effort, titled “Standardization of a Developmental Milestone Scale Using Data From Children in Israel“, has been published in JAMA Network Open.
Routine developmental screening tests for children are an important tool for timely identification of developmental delays. However, current assessment of developmental milestones lacks strong normative data. Using a large cohort of 643,958 children from a nationwide screening program, we established data-driven norms of milestone attainment and built a contemporary developmental scale, denoted THIS (Tipat Halav Israel Screening).
This work was a fruitful collaboration of researchers from KI Research Institute, the Israeli Ministry of Health and Ben-Gurion University, led by Dr. Yair Sadaka.
The THIS scale is on its way to be implemented in the Israeli routine developmental screening program, and is recommended for further evaluation worldwide.
As we face the 5th wave of the COVID-19 pandemic in Israel, we have focused on answering the question: what is the mental price of the pandemic? In our latest study we addressed this urging matter and compared the mental health of Israeli adolescents before and during the pandemic. We analyzed a nationwide health records data of over 200,000 (12-17 years old) from Maccabi Healthcare Services and calculated the incidence of mental health diagnoses and prescribed psychiatric drugs. Our research shows a significant increase (up to 55%) in all mental health diagnoses and most psychiatric drugs dispensation (up to 28%) during the COVID-19 period compared to a corresponding pre-COVID period. Most of the increase was associated with adolescent girls coming from the general Jewish sector. We believe that these results are just a “tip of the iceberg” as we are measuring only those who receive treatment in their HMOs while many teenagers are being treated in private clinics and their diagnoses and drug treatments are not recorded.
As we continue to explore these disturbing findings we advocate for continuous consideration of the mental wellbeing among adolescents while deciding on measures to mitigate the pandemic.Read Less
Our research on estimating the risk of postpartum depression from electronic health records using machine learning has been recently published in BMC Pregnancy and Childbirth. This research, which is part of KI’s focus on pregnancy complications, was done in collaboration with Dr. Jyotishman Pathak and Dr. Yiye Zhang from Weill Cornell Medicine, NY and with Dr. Vered Bar, Director of the Chava women’s mental health clinic at Sheba medical center.
Postpartum depression (PPD) is one of the most common complications of childbearing, estimated to affect 10-15% of mothers worldwide. In this research, we developed prediction models for identifying women at risk of PPD, using primary care electronic health records of over 260K pregnant women. Our machine learning-based models achieved good prediction performance and provided additive value to existing screening tools, which are based on symptom questionnaires. Furthermore, it allowed early alert of PPD risk prior to pregnancy. Incorporation of such models in the workflow of PPD screening may improve the subjectivity and accuracy of the screening process, enable timely interventions and consequently contribute to improved outcomes for the mother and child.
This has been such an exciting week at KI research institute as the pivotal manuscript describing the efficiency of the 3rd vaccine boost in Israel was published in the New England Journal of Medicine. This work proved the incredible value a true partnership and exceptional professionalism can bring – immediate impact on health in Israel and worldwide. We quickly realized that letting several research groups work in parallel on this invaluable data, using different approaches, giving mutual feedback and carefully analyzing this data with its inherited biases is the right way to advance knowledge regarding the vaccines efficiency. This research showed that the booster decreased confirmed infection in vaccinated individuals above 60 years of age by a factor of 11.3 and the rate of severe illness by a factor of 19.5. These insights could further support policy makers in determining whether to administer vaccine boost to the public during the global efforts to contain COVID-19. The study results were presented to the FDA advisory committee, who voted to recommend giving a booster dose to individuals 65 and older, adults above 16 at high risk for severe COVID-19 and those working in high risk settings.Read Less
We will be presenting two posters at the upcoming OHDSI Symposium (to be held virtually on Sep 14-15). In the first poster (more details here), we apply NLP methods on Wikipedia drug articles to identify adverse drug events which are frequently encountered in practice. And in the second poster (more details here), we address the challenge of estimating the performance of a trained prediction model on external datasets using only their limited statistical characteristics. Come visit our poster or contact us if you’re interested to hear more or want to join forces.Read Less
Breakthrough infections in vaccinated populations – for how long are we protected? KI research institute together with KSM research and innovation analyzed the new surge of COVID-19 cases in June 2021 which was composed mainly of the Delta (B.1.617.2) variant and included fully vaccinated individuals that received a two-dose regimen of the BioNTech/Pfizer mRNA BNT162b2 vaccine. To assess concerns of reduced vaccine efficacy as a result of limited duration of protection, we conducted a retrospective cohort study comparing the incidence rates of breakthrough infections between early and late vaccinees, using data from Maccabi Healthcare Services (MHS) containing 2.5 million members. The study population included fully vaccinated MHS members older than 16 years who received the second dose of the vaccine between January and April 2021. We applied two logistic regression models on matched populations where the outcome was defined as a positive SARS-CoV-2 PCR test recorded between June 1st and July 27th. We identified a significant correlation between time-from-vaccine and afforded protection against SARS-CoV-2 infection. The risk for breakthrough infection was significantly higher for early vaccinees compared to those vaccinated later.Read Less
Evaluating vaccine effectiveness upon high-risk exposure bears important implications as to quarantine policy for vaccinated individuals with confirmed SARS-CoV-2 exposure and protective guidelines in pandemic-ridden areas. We explored this question in a new collaborative study with KSM research and innovation center.
The uniqueness of this work stems from the scenario we studied: we examined the vaccine effectiveness in individuals with confirmed COVID-19 exposure – infected/exposed individuals who live in the same household. Since in this scenario all household members had a very high probability of being tested, we were also able to analyze asymptomatic infected individuals. In addition, since we included in our analysis vaccinated, unvaccinated and “recently vaccinated” (individuals exposed within 7 days of the first dose), we were able to control for other behaviors associated with receiving the vaccine.
The infection rates in household members (within 1-10 days of exposure) were 7.5% in fully vaccinated individuals, compared to 37.5% among unvaccinated and 41.7% among individuals recently vaccinated. This corresponded to vaccine effectiveness rates of 80.0% and 82.0% (fully vaccinated compared to unvaccinated or recently vaccinated).
The probabilities of testing for household member were 79.4%, 85.3% and 57.2% in the three vaccination groups (unvaccinated, recently vaccinated, fully vaccinated). These high testing rates enabled us to estimate the vaccine effectiveness in patients that are not necessarily in severe illness and are often asymptomatic. To our best knowledge, this is the first time such a real-world based analysis of vaccine effectiveness is reported.
Our observations suggest that household members of COVID-19 patients (and any individual with a confirmed exposure to COVID-19) are still at a considerable risk of being infected even if fully vaccinated and should get tested and use personal protection gear whenever possible.Read Less
Identifying patients at increased risk for severe COVID-19 is of high priority during the current pandemic. In our new study we assessed whether a second PCR test conducted during the first week after a SARS-CoV-2 positive test could identify patients at risk for developing severe illness.
In this study we analyzed data from electronic health records of 15,822 SARS-CoV-2 positive individuals and showed that a second negative PCR test result was associated with lower risk for severe illness compared to a positive result. The association was seen across different age groups and settings (community vs hospital). Importantly, this association was not limited to recovering patients but also observed in patients who still had evidence of COVID-19 as determined by a subsequent positive PCR test.
Our study suggests that multiple SARS-CoV-2 PCR testing may be used as an early, wide-spread complementary tool for risk assessment and subsequent appropriate disease management. Applying this methodology in COVID-19 clinical settings may alert physicians of patients at deterioration risk and facilitate early interventions, extend/limit isolation and allow better patient monitoring in hospital or at home.Read Less
Our latest research on symptoms dynamics of COVID-19, led by Barak Mizrahi, Nir Kalkstein and Karni Markus was published in Nature Communications this week.Read Less
This study was done in collaboration with researchers from Weizmann institute and Maccabi Institute for Research & Innovation. We studied over 200,000 people, among them 2471 tested positive for SARS-CoV-2 infection, and mapped reported symptoms before and after testing. Patients presented large variations in disease severity and recovery time with some symptoms lasting weeks. The most distinguishing factors for testing positive were loss of taste or smell (which was evident 3 weeks prior to testing) and fatigue. The study also included over 20,000 children, among them 862 tested positive and presented a faster recovery time compared to adults.
This study may assist in early identification of clinical symptoms for COVID-19, aid in the differential diagnosis, alert physicians for possible infection and facilitate timely testing, social isolation and treatment.
Our latest manuscript, with Maccabi Institute for Research & Innovation identified risk factors for COVID-19 complications. We studied a cohort of all SARS-CoV-2 positive individuals in a nationwide health organization (covering 2.3 million individuals) and identified those who suffered from serious complications; and compared the prevalence of pre-existing conditions, extracted from electronic health records, between complicated and non-complicated COVID-19 patient cohorts. Our analysis suggests that cardiovascular and kidney diseases, obesity, and hypertension are significant risk factors for COVID-19 complications, as previously reported. Interestingly, it also indicates that depression, as well as cognitive and neurological disorders, are significant risk factors; and that smoking and background of respiratory diseases do not significantly increase the risk of complications.Read Less