The Karma system is currently undergoing maintenance (Monday, January 29, 2018).
The maintenance period has been extended to 8PM EST.

Karma Credits will not be available for redeeming during maintenance.

Journal Description

JMIR Public Health & Surveillance (JPHS, Editor-in-chief: Travis Sanchez, Emory University/Rollins School of Public Health) is a PubMed-indexed, peer-reviewed sister journal of the Journal of Medical Internet Research (JMIR), the top cited journal in health informatics, ranked #1 by Clarivate's Journal Impact Factor. JPH is a multidisciplinary journal with a unique focus on the intersection of innovation and technology in public health, and includes topics like health communication, public health informatics, surveillance, participatory epidemiology, infodemiology and infoveillance, digital disease detection, digital public health interventions, mass media/social media campaigns, and emerging population health analysis systems and tools. 

We publish regular articles, reviews, protocols/system descriptions and viewpoint papers on all aspects of public health, with a focus on innovation and technology in public health.

Apart from publishing traditional public health research and viewpoint papers as well as reports from traditional surveillance systems, JPH was one of the first (if not the only) peer-reviewed journal which publishes papers with surveillance or pharmacovigilance data from non-traditional, unstructured big data and text sources such as social media and the Internet (infoveillance, digital disease detection), or reports on novel participatory epidemiology projects, where observations are solicited from the public.  

Among other innovations, JPH is also dedicated to support rapid open data sharing and rapid open access to surveillance and outbreak data. As one of the novel features we plan to publish rapid or even real-time surveillance reports and open data. The methods and description of the surveillance system may be peer-reviewed and published only once in detail, in a  "baseline report" (in a JMIR Res Protoc or a JMIR Public Health & Surveill paper), and authors then have the possibility to publish data and reports in frequent intervals rapidly and with only minimal additional peer-review (we call this article type "Rapid Surveillance Reports"). JMIR Publications may even work with authors/researchers and developers of selected surveillance systems on APIs for semi-automated reports (e.g. weekly reports to be automatically published in JPHS and indexed in PubMed, based on data-feeds from surveillance systems and minmal narratives and abstracts).

Furthermore, duing epidemics and public health emergencies, submissions with critical data will be processed with expedited peer-review to enable publication within days or even in real-time.

We also publish descriptions of open data resources and open source software. Where possible, we can and want to publish or even host the actual software or dataset on the journal website.


Recent Articles:

  • Source: Flickr; Copyright: GbergT; URL:; License: Creative Commons Attribution (CC-BY).

    Viral Loads Within 6 Weeks After Diagnosis of HIV Infection in Early and Later Stages: Observational Study Using National Surveillance Data


    Background: Early (including acute) HIV infection is associated with viral loads higher than those in later stages. Objective: This study aimed to examine the association between acute infection and viral loads near the time of diagnosis using data reported to the US National HIV Surveillance System. Methods: We analyzed data on infections diagnosed in 2012-2016 and reported through December 2017. Diagnosis and staging were based on the 2014 US surveillance case definition for HIV infection. We divided early HIV-1 infection (stage 0) into two subcategories. Subcategory 0α: a negative or indeterminate HIV-1 antibody test was ≤60 days after the first confirmed positive HIV-1 test or a negative or indeterminate antibody test or qualitative HIV-1 nucleic acid test (NAT) was ≤180 days before the first positive test, the latter being a NAT or detectable viral load. Subcategory 0β: a negative or indeterminate antibody or qualitative NAT was ≤180 days before the first positive test, the latter being an HIV antibody or antigen/antibody test. We compared median earliest viral loads for each stage and subcategory in each of the first 6 weeks after diagnosis using only the earliest viral load for each individual. Results: Of 203,392 infections, 56.69% (115,297/203,392) were reported with a quantified earliest viral load within 6 weeks after diagnosis and criteria sufficient to determine the stage at diagnosis. Among 5081 infections at stage 0, the median earliest viral load fell from 694,000 copies/mL in week 1 to 125,022 in week 2 and 43,473 by week 6. Among 30,910 infections in stage 1, the median earliest viral load ranged 15,412-17,495. Among 42,784 infections in stage 2, the median viral load declined from 44,973 in week 1 to 38,497 in week 6. Among 36,522 infections in stage 3 (AIDS), the median viral load dropped from 205,862 in week 1 to 119,000 in week 6. The median earliest viral load in stage 0 subcategory 0α fell from 1,344,590 copies/mL in week 1 to 362,467 in week 2 and 47,320 in week 6, while that in subcategory 0β was 70,114 copies/mL in week 1 and then 32,033 to 44,067 in weeks 2-6. The median viral load in subcategory 0α was higher than that in subcategory 0β in each of the first 6 weeks after diagnosis (P<.001). Conclusions: In the 1st week after diagnosis, viral loads in early infections are generally several times higher than those in later stages at diagnosis. By the 3rd week, however, most are lower than those in stage 3. High viral loads in early infection are much more common in subcategory 0α than in subcategory 0β, consistent with 0α comprising mostly acute infections and 0β comprising mostly postacute early infections. These findings may inform the prioritization of interventions for prevention.

  • M-Spot dried blood spot collection. Source: The Authors; Copyright: The Authors; URL:; License: Licensed by JMIR.

    Quantification of HIV-1 RNA Among Men Who Have Sex With Men Using an At-Home Self-Collected Dried Blood Spot Specimen: Feasibility Study


    Background: Suboptimal antiretroviral therapy (ART) adherence and disengagement in care present significant public health challenges because of the increased probability of HIV transmission. In the United States, men who have sex with men (MSM) continue to be disproportionately affected by HIV, highlighting a critical need to engage high-risk MSM living with HIV who are not engaged or retained in care. Objective: The aim of the study was to assess the feasibility of at-home blood self-collection and laboratory quantification of HIV-1 RNA viral load (VL) to report laboratory-based VL outcomes and compare self-reported and laboratory-reported VL Methods: Between 2016 and 2017, 766 US HIV-positive MSM enrolled in a Web-based behavioral intervention were invited to participate in an at-home dried blood spot (DBS) collection study using HemaSpot-HF kits (Spot On Sciences, Inc, Austin, TX) for laboratory-quantified VL. Results: Of those invited to participate, 72.3% (554/766) enrolled in the DBS study. Most (79.2%, 439/554) men enrolled reported attempting to collect their blood, 75.5% (418/554) of participants mailed a DBS specimen to the research laboratory, and 60.8% (337/554) had an adequate blood sample for VL testing. Of the 337 specimens tested for VL by the laboratory, 52.5% (177/337) had detectable VL (median: 3508 copies/mL; range: 851-1,202,265 copies/mL). Most men (83.9%, 135/161) who returned a DBS specimen with laboratory-quantified detectable VL self-reported an undetectable VL during their last clinical visit. Conclusions: Home collection of DBS samples from HIV-positive MSM is feasible and has the potential to support clinical VL monitoring. Discrepant laboratory HIV-1 RNA values and self-reported VL indicate a need to address perceived VL status, especially in the era of treatment as prevention. Most participants were willing to use an at-home DBS kit in the future, signaling an opportunity to engage high-risk MSM in long-term HIV care activities.

  • A surveillance officer using a mobile health tool FOR OUTBREAK RESPONSE. Source: Image created by the Authors; Copyright: SORMAS / Helmholtz Centre for Infection Research (HZI); URL:; License: Creative Commons Attribution (CC-BY).

    Assessing the Concepts and Designs of 58 Mobile Apps for the Management of the 2014-2015 West Africa Ebola Outbreak: Systematic Review


    Background: The use of mobile phone information technology (IT) in the health sector has received much attention especially during the 2014-2015 Ebola virus disease (EVD) outbreak. mHealth can be attributed to a major improvement in EVD control, but there lacks an overview of what kinds of tools were available and used based on the functionalities they offer. Objective: We aimed to conduct a systematic review of mHealth tools in the context of the recent EVD outbreak to identify the most promising approaches and guide further mHealth developments for infectious disease control. Methods: Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we searched for all reports on mHealth tools developed in the context of the 2014-2015 EVD outbreak published between January 1, 2014 and December 31, 2015 on Google Scholar, MEDLINE, CAB Abstracts (Global Health), POPLINE, and Web of Science in any language using the search strategy: (“outbreak” OR “epidemic”) AND (“mobile phone” OR “smartphone” OR “smart phone” OR “mobile phone” OR “tablet” OR “mHealth”) AND (“Ebola” OR ”EVD” OR “VHF” OR “Ebola virus disease” OR “viral hemorrhagic fever”) AND (“2014” OR “2015”). The relevant publications were selected by 2 independent reviewers who applied a standardized data extraction form on the tools’ functionalities. Results: We identified 1220 publications through the search strategy, of which 6.31% (77/1220) were original publications reporting on 58 specific mHealth tools in the context of the EVD outbreak. Of these, 62% (34/55) offered functionalities for surveillance, 22% (10/45) for case management, 18% (7/38) for contact tracing, and 6% (3/51) for laboratory data management. Only 3 tools, namely Community Care, Sense Ebola Followup, and Surveillance and Outbreak Response Management and Analysis System supported all four of these functionalities. Conclusions: Among the 58 identified tools related to EVD management in 2014 and 2015, only 3 appeared to contain all 4 key functionalities relevant for the response to EVD outbreaks and may be most promising for further development.

  • Source: (Jacob Sippel); Copyright: US Navy; URL:; License: Public Domain (CC0).

    Factors Related to Prostate-Specific Antigen–Based Prostate Cancer Screening in Primary Care: Retrospective Cohort Study of 120,587 French Men Over the Age...


    Background: International guidelines recommend avoiding prostate-specific antigen (PSA)-based prostate cancer screening in the elderly when life expectancy is less than 10 years. For younger men, most recommendations encourage a shared decision-making process taking into account patient comorbidities. Objective: The objective was to assess the performance of PSA-based prostate cancer screening in men older than 74 years and assess whether the presence (vs absence) of comorbidities was related to the performance of PSA testing in younger men aged 50 to 74 years who were eligible for screening. Methods: We analyzed data from the French national health care database (Loire-Atlantique geographic area). We reported the follow-up of two cohorts of men from April 1, 2014, to March 31, 2016: 22,480 men aged over 74 years and 98,107 men aged 50 to 74 years. We analyzed whether these patients underwent PSA testing after 2 years of follow-up and whether PSA testing performance was related to the following patient-related variables: age, low income, proxy measures indicative of major comorbidities (repeated ambulance transportation, having one of 30 chronic diseases, taking 5 or more drugs per day), or proxy measures indicative of specific comorbidities (cancer diseases, cardiovascular diseases, or psychiatric disorders). Statistical analysis was based on a multivariate mixed-effects logistic regression. Results: The proportion of patients who underwent a PSA-based screening test was 41.35% (9296/22,480) among men older than 74 years versus 41.05% (40,275/98,107) among men aged 50 to 74 years. The following factors were associated with less frequent PSA testing in men older than 74 years—age (odds ratio [OR] 0.89, 95% CI 0.88-0.89), low income (OR 0.18, 95% CI 0.05-0.69), suffering from a chronic disease (OR 0.82, 95% CI 0.76-0.88), repeated ambulance transportation (OR 0.37, 95% CI 0.31-0.44), diabetes requiring insulin (OR 0.51, 95% CI 0.43-0.60), dementia (OR 0.68, 95% CI 0.55-0.84), and antipsychotic treatment (OR 0.62, 95% CI 0.51-0.75)—whereas cardiovascular drug treatment was associated with more frequent PSA testing (OR 1.6, 95% CI 1.53-1.84). The following factors were associated with less frequent PSA testing in men aged 50 to 74 years—low income (OR 0.61, 95% CI 0.55-0.68); nonspecific conditions related to frailty: suffering from a chronic disease (OR 0.80, 95% CI 0.76-0.83), repeated ambulance transportation (OR 0.29, 95% CI 0.23-0.38), or chronic treatment with 5 or more drugs (OR 0.89, 95% CI 0.83-0.96); and various specific comorbidities: anticancer drug treatment (OR 0.67, 95% CI 0.55-0.83), diabetes requiring insulin (OR 0.55, 95% CI 0.49-0.61), and antiaggregant treatment (OR 0.91, 95% CI 0.86-0.96)—whereas older age (OR 1.07, 95% CI 1.07-1.08) and treatment with other cardiovascular drugs (OR 2.23, 95% CI 2.15-2.32) were associated with more frequent PSA testing. Conclusions: In this study, 41.35% (9296/22,480) of French men older than 74 years had a PSA-based screening test. Although it depends on patient comorbidities, PSA testing remains inappropriate in certain populations.

  • Source: Pexels; Copyright: LinkedIn Sales Navigator; URL:; License: Licensed by the authors.

    Quality of HIV Websites With Information About Pre-Exposure Prophylaxis or Treatment as Prevention for Men Who Have Sex With Men: Systematic Evaluation


    Background: Knowledge and uptake of high-efficacy HIV prevention strategies such as pre-exposure prophylaxis (PrEP) and treatment as prevention (TasP) remain low among men who have sex with men (MSM) who are at the highest risk for HIV infection in the United States. Electronic health (eHealth) interventions are promising tools for disseminating information about these critical yet underutilized strategies and addressing key barriers to uptake among target populations. However, existing HIV prevention websites are understudied and unevaluated. Objective: This study aimed to systematically review and evaluate existing HIV websites that include information about PrEP or TasP for MSM. Methods: From March 2018 to May 2018, 2 trained research assistants (RAs) entered relevant key words and phrases into 3 commonly used search engines and applied exclusion criteria to all returned results to identify 31 websites included in this review. RAs independently scored each website for authority, usability, interactivity, and PrEP/TasP-related content based on a standardized rating scale and then averaged the results. Results: No website received a perfect score in any of the 4 categories, and the average website score was 62% (37/60). Less than a quarter of the websites (23%, 7/31) received a score of more than 75% (7.5/10) for content. Approximately two-thirds of the websites (65%, 20/31) received a score of 50% (5/10) or lower for interactivity. The average score in usability was 68% (6.8/10) and in authority was 69% (6.9/10). Other deficiencies observed included difficulty locating relevant content and lack of information targeting audiences with the highest likelihood of HIV infection. Conclusions: Existing HIV prevention websites with information about PrEP or TasP for MSM fail to provide adequate content as well as present that content to users in an interactive and audience-conscious way. Future eHealth interventions should attempt to rectify these deficiencies to successfully engage and educate MSM at high risk for HIV regarding prevention strategies.

  • Influenzanet website (montage). Source: The Authors; Copyright: The Authors; URL:; License: Licensed by JMIR.

    Research Ethics in the European Influenzanet Consortium: Scoping Review


    Background: Influenzanet was launched in several European countries to monitor influenza-like illness during flu seasons with the help of volunteering participants and Web-based technologies. As in the case of developing fields, ethical approaches are not well developed in the collection, processing, and analysis of participants’ information. Existing controversies and varying national ethical regulations can, thus, hamper efficient cross-border research collaboration to the detriment of quality disease surveillance. Objective: This scoping review characterizes current practices on how ethical, legal, and social issues (ELSIs) pertinent to research ethics are handled by different Influenzanet country groups to analyze similarities and identify the need for further harmonization of ethical approaches. Methods: A literature search was carried out on PubMed, Web of Science, Global Digital Library on Ethics, and Bioethics Literature Database to identify ELSIs for Influenzanet country platforms. Only English-language papers were included with publication dates from 2003 to 2017. Publications were screened for the application of bioethics principles in the implementation of country platforms. Additional publications gathered from the Influenzanet Consortium website, reference screening, and conference proceeding were screened for ELSIs. Results: We gathered 96 papers from our search methodology. In total, 28 papers that mentioned ELSIs were identified and included in this study. The Research Ethics Committee (REC) approvals were sought for recruiting participants and collecting their data in 8 of 11 country platforms and informed e-consent was sought from participants in 9 of 11 country platforms. Furthermore, personal data protection was ensured throughout the Consortium using data anonymization before processing and analysis and using aggregated data. Conclusions: Epidemics forecasting activities, such as Influenzanet, are beneficial; however, its benefits could be further increased through the harmonization of data gathering and ethical requirements. This objective is achievable by the Consortium. More transparency should be promoted concerning REC-approved research for Influenzanet-like systems. The validity of informed e-consent could also be increased through the provision of a user friendly and standard information sheet across the Consortium where participants agree to its terms, conditions, and privacy policies before being able to fill in the questionnaire. This will help to build trust in the general public while preventing any decline in participation.

  • SOMAARTH Demographic, Development and Environmental Surveillance Site (montage). Source:; Copyright: The INCLEN Trust International; URL:; License: Licensed by JMIR.

    Establishing a Demographic, Development and Environmental Geospatial Surveillance Platform in India: Planning and Implementation


    Background: Inadequate administrative health data, suboptimal public health infrastructure, rapid and unplanned urbanization, environmental degradation, and poor penetration of information technology make the tracking of health and well-being of populations and their social determinants in the developing countries challenging. Technology-integrated comprehensive surveillance platforms have the potential to overcome these gaps. Objective: This paper provides methodological insights into establishing a geographic information system (GIS)-integrated, comprehensive surveillance platform in rural North India, a resource-constrained setting. Methods: The International Clinical Epidemiology Network Trust International established a comprehensive SOMAARTH Demographic, Development, and Environmental Surveillance Site (DDESS) in rural Palwal, a district in Haryana, North India. The surveillance platform evolved by adopting four major steps: (1) site preparation, (2) data construction, (3) data quality assurance, and (4) data update and maintenance system. Arc GIS 10.3 and QGIS 2.14 software were employed for geospatial data construction. Surveillance data architecture was built upon the geospatial land parcel datasets. Dedicated software (SOMAARTH-1) was developed for handling high volume of longitudinal datasets. The built infrastructure data pertaining to land use, water bodies, roads, railways, community trails, landmarks, water, sanitation and food environment, weather and air quality, and demographic characteristics were constructed in a relational manner. Results: The comprehensive surveillance platform encompassed a population of 0.2 million individuals residing in 51 villages over a land mass of 251.7 sq km having 32,662 households and 19,260 nonresidential features (cattle shed, shops, health, education, banking, religious institutions, etc). All land parcels were assigned georeferenced location identification numbers to enable space and time monitoring. Subdivision of villages into sectors helped identify socially homogenous community clusters (418/676, 61.8%, sectors). Water and hygiene parameters of the whole area were mapped on the GIS platform and quantified. Risk of physical exposure to harmful environment (poor water and sanitation indicators) was significantly associated with the caste of individual household (P=.001), and the path was mediated through the socioeconomic status and density of waste spots (liquid and solid) of the sector in which these households were located. Ground-truthing for ascertaining the land parcel level accuracies, community involvement in mapping exercise, and identification of small habitations not recorded in the administrative data were key learnings. Conclusions: The SOMAARTH DDESS experience allowed us to document and explore dynamic relationships, associations, and pathways across multiple levels of the system (ie, individual, household, neighborhood, and village) through a geospatial interface. This could be used for characterization and monitoring of a wide range of proximal and distal determinants of health.

  • Japanese businessmen wearing masks to avoid getting the flu in the city. Source: iStock by Getty Images; Copyright: franckreporter; URL:; License: Licensed by the authors.

    Twitter-Based Influenza Detection After Flu Peak via Tweets With Indirect Information: Text Mining Study


    Background: The recent rise in popularity and scale of social networking services (SNSs) has resulted in an increasing need for SNS-based information extraction systems. A popular application of SNS data is health surveillance for predicting an outbreak of epidemics by detecting diseases from text messages posted on SNS platforms. Such applications share the following logic: they incorporate SNS users as social sensors. These social sensor–based approaches also share a common problem: SNS-based surveillance are much more reliable if sufficient numbers of users are active, and small or inactive populations produce inconsistent results. Objective: This study proposes a novel approach to estimate the trend of patient numbers using indirect information covering both urban areas and rural areas within the posts. Methods: We presented a TRAP model by embedding both direct information and indirect information. A collection of tweets spanning 3 years (7 million influenza-related tweets in Japanese) was used to evaluate the model. Both direct information and indirect information that mention other places were used. As indirect information is less reliable (too noisy or too old) than direct information, the indirect information data were not used directly and were considered as inhibiting direct information. For example, when indirect information appeared often, it was considered as signifying that everyone already had a known disease, leading to a small amount of direct information. Results: The estimation performance of our approach was evaluated using the correlation coefficient between the number of influenza cases as the gold standard values and the estimated values by the proposed models. The results revealed that the baseline model (BASELINE+NLP) shows .36 and that the proposed model (TRAP+NLP) improved the accuracy (.70, +.34 points). Conclusions: The proposed approach by which the indirect information inhibits direct information exhibited improved estimation performance not only in rural cities but also in urban cities, which demonstrated the effectiveness of the proposed method consisting of a TRAP model and natural language processing (NLP) classification.

  • Facebook advertisement posted to Canadian parents’ News Feeds (montage). Source: The Authors /; Copyright: JMIR Publications; URL:; License: Creative Commons Attribution (CC-BY).

    User-Driven Comments on a Facebook Advertisement Recruiting Canadian Parents in a Study on Immunization: Content Analysis


    Background: More people are searching for immunization information online and potentially being exposed to misinformation and antivaccination sentiment in content and discussions on social media platforms. As vaccination coverage rates remain suboptimal in several developed countries, and outbreaks of vaccine-preventable diseases become more prevalent, it is important that we build on previous research by analyzing themes in online vaccination discussions, including those that individuals may see without actively searching for information on immunization. Objective: The study aimed to explore the sentiments and themes behind an unsolicited debate on immunization in order to better inform public health interventions countering antivaccination sentiment. Methods: We analyzed and quantified 117 user-driven open-ended comments on immunization posted in the Comments section of a Facebook advertisement that targeted Canadian parents for recruitment into a larger study on immunization. Then, 2 raters coded all comments using content analysis. Results: Of 117 comments, 85 were posted by unique commentators, with most being female (65/85, 77%). The largest proportion of the immunization comments were positive (51/117, 43.6%), followed by negative (41/117, 35.0%), ambiguous (20/117, 17.1%), and hesitant (5/117, 4.3%). Inaccurate knowledge (27/130, 20.8%) and misperceptions of risk (23/130, 17.7%) were most prevalent in the 130 nonpositive comments. Other claims included distrust of pharmaceutical companies or government agencies (18/130, 13.8%), distrust of the health care system or providers (15/130, 11.5%), past negative experiences with vaccination or beliefs (10/130, 7.7%), and attitudes about health and prevention (10/130, 7.7%). Almost 40% (29/74, 39%) of the positive comments communicated the risks of not vaccinating, followed by judgments on the knowledge level of nonvaccinators (13/74, 18%). A total of 10 positive comments (10/74, 14%) specifically refuted the link between autism and vaccination. Conclusions: The presence of more than 100 unsolicited user-driven comments on a platform not intended for discussion, nor providing any information on immunization, illustrates the strong sentiments associated with immunization and the arbitrariness of the online platforms used for immunization debates. Health authorities should be more proactive in finding mechanisms to refute misinformation and misperceptions that are propagating uncontested online. Online debates and communications on immunization need to be identified by continuous monitoring in order for health authorities to understand the current themes and trends, and to engage in the discussion.

  • Source: Getty Images; Copyright: stevanovicigor; URL:; License: Licensed by the authors.

    eHealth Literacy in People Living with HIV: Systematic Review


    Background: In the era of eHealth, eHealth literacy is emerging as a key concept to promote self-management of chronic conditions such as HIV. However, there is a paucity of research focused on eHealth literacy for people living with HIV (PLWH) as a means of improving their adherence to HIV care and health outcome. Objective: The objective of this study was to critically appraise the types, scope, and nature of studies addressing eHealth literacy as a study variable in PLWH. Methods: This systematic review used comprehensive database searches, such as PubMed, EMBASE, CINAHL, Web of Science, and Cochrane, to identify quantitative studies targeting PLWH published in English before May 2017 with eHealth literacy as a study variable. Results: We identified 56 unique records, and 7 papers met the eligibility criteria. The types of study designs varied (descriptive, n=3; quasi-experimental, n=3; and experimental, n=1) and often involved community-based settings (n=5), with sample sizes ranging from 18 to 895. In regards to instruments used, 3 studies measured eHealth literacy with validated instruments such as the eHealth Literacy Scale (eHEALS); 2 studies used full or short versions of Test of Functional Health Literacy in Adults, whereas the remaining 2 studies used study-developed questions. The majority of studies included in the review reported high eHealth literacy among the samples. The associations between eHealth literacy and health outcomes in PLWH were not consistent. In the areas of HIV transmission risk, retention in care, treatment adherence, and virological suppression, the role of eHealth literacy is still not fully understood. Furthermore, the implications for future research are discussed. Conclusions: Understanding the role of eHealth literacy is an essential step to encourage PLWH to be actively engaged in their health care. Avenues to pursue in the role of eHealth literacy and PLWH should consider the development and use of standardized eHealth literacy definitions and measures.

  • Source: Image created by the Authors; Copyright: Interactive Research and Development; URL:; License: Creative Commons Attribution (CC-BY).

    Using Predictive Analytics to Identify Children at High Risk of Defaulting From a Routine Immunization Program: Feasibility Study


    Background: Despite the availability of free routine immunizations in low- and middle-income countries, many children are not completely vaccinated, vaccinated late for age, or drop out from the course of the immunization schedule. Without the technology to model and visualize risk of large datasets, vaccinators and policy makers are unable to identify target groups and individuals at high risk of dropping out; thus default rates remain high, preventing universal immunization coverage. Predictive analytics algorithm leverages artificial intelligence and uses statistical modeling, machine learning, and multidimensional data mining to accurately identify children who are most likely to delay or miss their follow-up immunization visits. Objective: This study aimed to conduct feasibility testing and validation of a predictive analytics algorithm to identify the children who are likely to default on subsequent immunization visits for any vaccine included in the routine immunization schedule. Methods: The algorithm was developed using 47,554 longitudinal immunization records, which were classified into the training and validation cohorts. Four machine learning models (random forest; recursive partitioning; support vector machines, SVMs; and C-forest) were used to generate the algorithm that predicts the likelihood of each child defaulting from the follow-up immunization visit. The following variables were used in the models as predictors of defaulting: gender of the child, language spoken at the child’s house, place of residence of the child (town or city), enrollment vaccine, timeliness of vaccination, enrolling staff (vaccinator or others), date of birth (accurate or estimated), and age group of the child. The models were encapsulated in the predictive engine, which identified the most appropriate method to use in a given case. Each of the models was assessed in terms of accuracy, precision (positive predictive value), sensitivity, specificity and negative predictive value, and area under the curve (AUC). Results: Out of 11,889 cases in the validation dataset, the random forest model correctly predicted 8994 cases, yielding 94.9% sensitivity and 54.9% specificity. The C-forest model, SVMs, and recursive partitioning models improved prediction by achieving 352, 376, and 389 correctly predicted cases, respectively, above the predictions made by the random forest model. All models had a C-statistic of 0.750 or above, whereas the highest statistic (AUC 0.791, 95% CI 0.784-0.798) was observed in the recursive partitioning algorithm. Conclusions: This feasibility study demonstrates that predictive analytics can accurately identify children who are at a higher risk for defaulting on follow-up immunization visits. Correct identification of potential defaulters opens a window for evidence-based targeted interventions in resource limited settings to achieve optimal immunization coverage and timeliness.

  • Source: Wikimedia Commons; Copyright: NIAID; URL:; License: Creative Commons Attribution (CC-BY).

    Overlap of Asthma and Chronic Obstructive Pulmonary Disease in Patients in the United States: Analysis of Prevalence, Features, and Subtypes


    Background: Although asthma and chronic obstructive pulmonary disease (COPD) are clinically distinct diseases, they represent biologically diverse and overlapping clinical entities and it has been observed that they often co-occur. Some research and theorizing suggest there is a common comorbid condition termed asthma-chronic obstructive pulmonary disease overlap (ACO). However, the existence of ACO is controversial. Objective: The objective of this study is to describe patient characteristics and estimate prevalence, health care utilization, and costs of ACO using claims-based diagnoses confirmed with medical record information. Methods: Eligible patients were commercial US health plan enrollees; ≥40 years; had asthma, COPD, or ACO; ≥3 prescription fills for asthma/COPD medications; and ≥2 spirometry tests. Records for a random sample of 5000 patients with ACO were reviewed to validate claims-based diagnoses. Results: The estimated ACO prevalence was 6% (estimated 10,250/183,521) among 183,521 full study patients. In the claims-based cohorts, the comorbidity burden for ACO was greater versus asthma but similar to COPD cohorts. Medication utilization was higher in ACO versus asthma and COPD. Mean total health care costs were significantly higher for ACO versus asthma but similar to COPD. In confirmed diagnoses cohorts, mean total health care costs (medical plus pharmacy) were lower for ACO versus COPD but similar to asthma (US $20,035; P=.56). Among confirmed cases, where there was medical record evidence, smoking history was higher in ACO (300/343, 87.5%) versus asthma cohorts (100/181, 55.2%) but similar to COPD (68/84, 81%). Conclusions: ACO had more comorbidities, medication utilization, and costs than patients with asthma or COPD but differences were not seen after confirmation with medical records.

Citing this Article

Right click to copy or hit: ctrl+c (cmd+c on mac)

Latest Submissions Open for Peer-Review:

View All Open Peer Review Articles
  • Using evidence from Programmatic Mapping approaches to design and implement high impact HIV prevention programs for female sex workers

    Date Submitted: Oct 29, 2018

    Open Peer Review Period: Nov 8, 2018 - Nov 22, 2018

    Programmatic Mapping (PM) is a rapid and efficient mechanism to develop size estimates of key populations (KPs) and geo-locate them at physical locations in a systematic and scientific manner. At the...

    Programmatic Mapping (PM) is a rapid and efficient mechanism to develop size estimates of key populations (KPs) and geo-locate them at physical locations in a systematic and scientific manner. At the macro level, this information forms the basis for allocating program resources, setting performance targets, assessing coverage, and determining funding requirements to saturate FSW coverage. At a micro level, PM data provide specific information on hot spots, estimates of FSWs at those spots, and hotspot typology and days and times of operation, all of which informs targeted service delivery strategies. This information can provide a reliable platform to plan HIV prevention and treatment services to considerable scale and intensity. Above all, the entire PM process requires deep involvement of FSWs, which increases community ownership of the data and can lead to an increased uptake of services. Despite a few limitations, the approach is versatile and can be used in varied country contexts to generate important information about sex work and its dynamics. In this paper, we describe experiences and lessons learned from using evidence generated from programmatic mapping of female sex workers (FSWs) in multiple countries to develop HIV prevention programs at scale.