On Measuring Vaccine Effectiveness with Observational Study Designs.

Herein, we present a bird's eye view of common observational study designs utilized for measurement of vaccine effectiveness. Assessing vaccines effectiveness is an integral part of vaccine research, particularly for the newly developed vaccines. A cohort study is prospective, directing from an exposure to one or more outcomes. The design is the best method to ascertain the attack rate of an infectious disease. A traditional case-control study is retrospective, directing from a given outcome to one or more exposures. The design cannot provide the relative risk, but it can provide the odds ratio, which is a good estimation of the relative risk when the attack rate is low. Critically depending on laboratory test results and performance, the test-negative case-control study design is another type of observational study commonly used nowadays for the evaluation of the vaccine effectiveness. Comparing to cohort and traditional case-control designs, conducting a test-negative case-control study is relatively cheaper and faster. Herein, we describe each of the above-mentioned study designs through examples generated by a Monte-Carlo simulation program assuming real-world conditions. CONCLUSION: The simulation shows that regardless of the study design employed, the diagnostic test specificity is of utmost importance in providing a valid estimate of the vaccine effectiveness.

Determining the SARS-CoV-2 serological immunoassay test performance indices based on the test results frequency distribution.

Introduction: Coronavirus disease 2019 (COVID-19) is known to induce robust antibody response in most of the affected individuals. The objective of the study was to determine if we can harvest the test sensitivity and specificity of a commercial serologic immunoassay merely based on the frequency distribution of the SARS-CoV-2 immunoglobulin (Ig) G concentrations measured in a population-based seroprevalence study. Materials and methods: The current study was conducted on a subset of a previously published dataset from the canton of Geneva. Data were taken from two non-consecutive weeks (774 samples from May 4-9, and 658 from June 1-6, 2020). Assuming that the frequency distribution of the measured SARS-CoV-2 IgG is binormal (an educated guess), using a non-linear regression, we decomposed the distribution into its two Gaussian components. Based on the obtained regression coefficients, we calculated the prevalence of SARS-CoV-2 infection, the sensitivity and specificity, and the most appropriate cut-off value for the test. The obtained results were compared with those obtained from a validity study and a seroprevalence population-based study. Results: The model could predict more than 90% of the variance observed in the SARS-CoV-2 IgG distribution. The results derived from our model were in good agreement with the results obtained from the seroprevalence and validity studies. Altogether 138 of 1432 people had SARS-CoV-2 IgG ≥ 0.90, the cut-off value which maximized the Youden's index. This translates into a true prevalence of 7.0% (95% confidence interval 5.4% to 8.6%), which is in keeping with the estimated prevalence of 7.7% derived from our model. Our model can provide the true prevalence. Conclusions: Having an educated guess about the distribution of test results, the test performance indices can be derived with acceptable accuracy merely based on the test results frequency distribution without the need for conducting a validity study and comparing the test results against a gold-standard test.

The apparent prevalence, the true prevalence.

Serologic tests are important for conducting seroepidemiologic and prevalence studies. However, the tests used are typically imperfect and produce false-positive and false-negative results. This is why the seropositive rate (apparent prevalence) does not typically reflect the true prevalence of the disease or condition of interest. Herein, we discuss the way the true prevalence could be derived from the apparent prevalence and test sensitivity and specificity. A computer simulation based on the Monte-Carlo algorithm was also used to further examine a situation where the measured test sensitivity and specificity are also uncertain. We then complete our review with a real example. The apparent prevalence observed in many prevalence studies published in medical literature is a biased estimation and cannot be interpreted correctly unless we correct the value.

Use of Oral Polio Vaccine and the Global Incidence of Mother-to-Child Human Immunodeficiency Virus Transmission.

Background: Mother-to-child transmission (MTCT) of human immunodeficiency virus (HIV) is an important global health issue. We hypothesized that the live attenuated poliovirus existing in oral polio vaccine (OPV) may protect uninfected neonates born to HIV-positive mothers through the stimulation of innate immune system. Objective: To test the hypothesis that countries using OPV have a lower MTCT rate (due to postnatal protection provided by the vaccine) compared with those using only inactivated polio vaccine (IPV). Methods: In an ecological study, the incidence of HIV/AIDS in children aged <1 year (IncHIV1), considered a surrogate index for MTCT rate, was compared between countries using OPV vs. IPV. The aggregated population data were retrieved for 204 countries from the Global Burden of Disease (GBD 2019) Collaborative Network website, "Our World in Data" website, the World Bank website, and the WHO Global Polio Eradication Initiative (GPEI). We used a negative binomial regression model with IncHIV1 as the dependent variable and the prevalence of HIV/AIDS in women aged 15-49 years (PrevHIV), antiretroviral therapy (ART) coverage, human development index (HDI), and the type of vaccine used in each country as independent variables. Multivariate imputation by chained equations was used to treat missing values. Analyses were performed for both the original dataset (with missing values) and the five imputed datasets. Results: IncHIV1 and PrevHIV were available for all 204 countries; vaccine type, 194 countries; HDI, 182 countries; and ART coverage, 133 countries. One-hundred and twenty-nine countries in the original dataset had complete data for all the above-mentioned variables; the imputed datasets had complete data for all 204 countries. The results obtained from the analysis of the original dataset had no overall difference with the pooled results obtained from the analysis of the five imputed datasets. Countries with higher HDI mainly use IPV; those with lower HDI commonly use OPV. PrevHIV, HDI, and the type of vaccine were independent predictors of IncHIV1. Use of OPV compared to IPV, was independently associated with an average decrease of 17% in IncHIV1 at the median HDI of 0.75. The protection provided by OPV increased in countries with lower HDI. Conclusions: Use of OPV compared with IPV, was independently associated with lower MTCT rate.

Use of oral polio vaccine and the incidence of COVID-19 in the world.

BACKGROUND: Several live attenuated vaccines were shown to provide temporary protection against a variety of infectious diseases through stimulation of the host innate immune system. OBJECTIVE: To test the hypothesis that countries using oral polio vaccine (OPV) have a lower cumulative number of cases diagnosed with COVID-19 per 100,000 population (CP100K) compared with those using only inactivated polio vaccine (IPV). METHODS: In an ecological study, the CP100K was compared between countries using OPV vs IPV. We used a random-effect meta-analysis technique to estimate the pooled mean for CP100K. We also used negative binomial regression with CP100K as the dependent variable and the human development index (HDI) and the type of vaccine used as independent variables. RESULTS: The pooled estimated mean CP100K was 4970 (95% CI 4030 to 5900) cases per 100,000 population for countries using IPV, significantly (p<0.001) higher than that for countries using OPV-1580 (1190 to 1960). Countries with higher HDI prefer to use IPV; those with lower HDI commonly use OPV. Both HDI and the type of vaccine were independent predictors of CP100K. Use of OPV compared to IPV could independently decrease the CP100K by an average of 30% at the mean HDI of 0.72. CONCLUSIONS: Countries using OPV have a lower incidence of COVID-19 compared to those using IPV. This might suggest that OPV may either prevent SARS-CoV-2 infection at individual level or slow down the transmission at the community level.

COVID-19 Infection Among Women in Iran Exposed vs Unexposed to Children Who Received Attenuated Poliovirus Used in Oral Polio Vaccine.

Importance: Live attenuated vaccines may provide short-term protection against infectious diseases through stimulation of the innate immune system. Objective: To evaluate whether passive exposure to live attenuated poliovirus is associated with diminished symptomatic infection with SARS-CoV-2. Design, Setting, and Participants: In a longitudinal cohort study involving 87 923 people conducted between March 20 and December 20, 2020, the incidence of COVID-19 was compared between 2 groups of aged-matched women with and without exposure to live attenuated poliovirus in the oral polio vaccine (OPV). Participants were people receiving health care services from the Petroleum Industry Health Organization and residing in 2 cities in Iran (ie, Ahwaz and Shiraz). Participants were women aged 18 to 48 years whose children were aged 18 months or younger and a group of age-matched women from the same residence who had had no potential exposure to OPV. Exposures: Indirect exposure to live attenuated poliovirus in OPV. Main Outcomes and Measures: Symptomatic COVID-19, diagnosed by reverse transcription-polymerase chain reaction. Results: After applying the inclusion and exclusion criteria, 419 mothers (mean [SD] age, 35.5 [4.9] years) indirectly exposed to the OPV and 3771 age-matched women (mean [SD] age, 35.7 [5.3] years) who had no exposure to OPV were available for analysis. COVID-19 was diagnosed in 1319 of the 87 923 individuals in the study population (151 per 10 000 population) during the study period. None of the mothers whose children received OPV developed COVID-19 after a median follow-up of 141 days (IQR, 92-188 days; range, 1-270 days); 28 women (0.74%; 95% CI, 0.47%-1.02%) in the unexposed group were diagnosed with COVID-19 during the 9 months of the study. Point-by-point comparison of the survival curves of the exposed and unexposed groups found that indirect exposure to OPV was significantly associated with decreased COVID-19 acquisition; probability of remaining without infection was 1.000 (95% CI, 1.000-1.000) in the exposed group vs 0.993 (95% CI, 0.990-0.995) in the unexposed group after 9 months (P < .001). Conclusions and Relevance: In this cohort study, indirect exposure to live attenuated poliovirus was associated with decreased symptomatic infection with COVID-19. Further study of the potential protective effect of OPV should be conducted, especially in nations where OPV is already in use for polio prevention and specific COVID-19 vaccines are delayed, less affordable, or fail to meet demand.

On the information hidden in a classifier distribution.

Classification tasks are a common challenge to every field of science. To correctly interpret the results provided by a classifier, we need to know the performance indices of the classifier including its sensitivity, specificity, the most appropriate cut-off value (for continuous classifiers), etc. Typically, several studies should be conducted to find all these indices. Herein, we show that they already exist, hidden in the distribution of the variable used to classify, and can readily be harvested. An educated guess about the distribution of the variable used to classify in each class would help us to decompose the frequency distribution of the variable in population into its components-the probability density function of the variable in each class. Based on the harvested parameters, we can then calculate the performance indices of the classifier. As a case study, we applied the technique to the relative frequency distribution of prostate-specific antigen, a biomarker commonly used in medicine for the diagnosis of prostate cancer. We used nonlinear curve fitting to decompose the variable relative frequency distribution into the probability density functions of the non-diseased and diseased people. The functions were then used to determine the performance indices of the classifier. Sensitivity, specificity, the most appropriate cut-off value, and likelihood ratios were calculated. The reference range of the biomarker and the prevalence of prostate cancer for various age groups were also calculated. The indices obtained were in good agreement with the values reported in previous studies. All these were done without being aware of the real health status of the individuals studied. The method is even applicable for conditions with no definite definitions (e.g., hypertension). We believe the method has a wide range of applications in many scientific fields.

Accessing free quality health care services and premature death.

BACKGROUND: The type of health insurance may affect the likelihood of mortality of insured people. We conducted this study to determine if accessing free quality health care services could decrease the premature mortality of people in a developing country. METHODS: In a multicenter cross sectional study, "years-life-lost" (YLL) due to premature death was evaluated in 202 671 insured people residing in six large regions in Iran. The participants had access to free quality health care services. The number of insured people that died in the six regions during March 20, 2014, to March 20, 2015, as well as their sex, age, and cause of the death, were collected, and the YLL was calculated based on assumptions made in Global Burden of Disease Study 2010 (GBD2010). RESULTS: The crude mortality rate was 2.3 per 1000, significantly lower than the overall rate of 4.6 per 1000 people in general population of Iran. The average YLL was 47 years per 1000 persons, significantly lower than that in general population of Iran and many industrialized countries. The most common causes of death (and YLL) were cardiovascular diseases and malignancies. CONCLUSION: Having access to free quality health care services is associated with a significant decrease in premature death.

On determining the most appropriate test cut-off value: the case of tests with continuous results.

There are several criteria for determination of the most appropriate cut-off value in a diagnostic test with continuous results. Mostly based on receiver operating characteristic (ROC) analysis, there are various methods to determine the test cut-off value. The most common criteria are the point on ROC curve where the sensitivity and specificity of the test are equal; the point on the curve with minimum distance from the left-upper corner of the unit square; and the point where the Youden's index is maximum. There are also methods mainly based on Bayesian decision analysis. Herein, we show that a proposed method that maximizes the weighted number needed to misdiagnose, an index of diagnostic test effectiveness we previously proposed, is the most appropriate technique compared to the aforementioned ones. For determination of the cut-off value, we need to know the pretest probability of the disease of interest as well as the costs incurred by misdiagnosis. This means that even for a certain diagnostic test, the cut-off value is not universal and should be determined for each region and for each disease condition.

Knowledge of and attitudes toward breast self-examination in Iranian women: a multi-center study.

BACKGROUND: In Iran, breast cancer is the second most common malignancy among women and diagnosed in advanced stages almost a decade earlier than women living in western countries. OBJECTIVE: To determine the level of knowledge and attitudes toward breast self-examination (BSE) among Iranian women. METHODS: In a cross-sectional multi-center study, 3,060 women aged between 25 and 54 years attending between July and October 2009 to 11 general health care centers located in 11 cities of Iran, were interviewed to assess their knowledge and attitudes toward BSE. RESULTS: Of the total, the data of 3,030 were found eligible for analyses. The median (interquartile range [IQR]) age of participants was 40 (14) years; 2,687 (88.7%) were married; 1496 (49.4%) women performed BSE, 290 of whom (19.4% of performers and 9.6% of all studied women) did it using a correct method and at an appropriate time. More than three-quarters of performers (n=1141) learned about BSE from a health care provider; 248 (16.6%, 95% confidence interval [CI]: 14.7% to 18.5%) had abnormal findings, 11 (0.7% , 95% CI: 0.3% to 1.2%) of which were malignant. Of 1,534 (50.6%) non-performers, 474 (30.9%) did not know how to do BSE; the remaining women did not do BSE for fear of being found positive for cancer or did not care about it. Being a health care provider, married, educated, and knowing of a person with breast cancer was associated with performing BSE. CONCLUSIONS: The level of BSE practice and knowledge among Iranian women is unsatisfactory. We should emphasize appropriate education of women.

Are shorter article titles more attractive for citations? Cross-sectional study of 22 scientific journals.

AIM: To investigate the correlation between the length of the title of a scientific article and the number of citations it receives, in view of the common editorial call for shorter titles. METHODS: Title and the number of citations to all articles published in 2005 in 22 arbitrarily chosen English-language journals (n=9031) were retrieved from citation database Scopus. The 2008 journal impact factors of these 22 journals were also retrieved from Thomson Reuters' Journal Citation Report (JCR). Assuming the article title length as the independent variable, and the number of citations to the article as the dependent variable, a linear regression model was applied. RESULTS: The slope of the regression line for some journals (n=6, when titles were measured in characters but 7 when titles were measured in words) was negative - none was significantly different from 0. The overall slope for all journals was 0.140 (when titles were measured in characters) and 0.778 (when titles were measured in words), which is significantly different from 0 (P<0.001 , t test). Overall, articles with longer titles received more citations - Spearman rho=0.266 - when titles were measured in characters, and rho=0.244 when titles were measured in words (P<0.001). This association was found for 7 of 8 journals with impact factor >10 and for 2 out of 14 journals with impact factor <10 (P<0.001, Fisher exact test). CONCLUSION: Longer titles seem to be associated with higher citation rates. This association is more pronounced for journals with high impact factors. Editors who insist on brief and concise titles should perhaps update the guidelines for authors of their journals and have more flexibility regarding the length of the title.