Nishant Kishore’s research while affiliated with Beverly Hospital, Boston MA and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (23)


Metapopulation metric distribution for different values of epsilon for Scenario A: Location of the first cases, B: change in the mobility network, and C: change in the parameters of the metapopulation model. As metrics can exist on very different scales, we calculate the normalized distribution of bootstrapped metrics where a minimum amount of noise is added. We then compare this to the median value of bootstrapped values at increasing values of noise to describe the change from expectation.
Flow diagram showing the architecture of the modular software pipeline designed to quantify the tradeoff privacy utility of mobility data post-differential privacy processing in epidemiological models
The boxes in pink represent the DP process, the boxes in red represent cleansing processes by both data providers and modelers, the boxes in green represent the data, the boxes in white are predefined processes, and the box in yellow stands for the metapopulation model.
Parameters of the metapopulation model
Overview of the scenarios and the investigated question
A standardised differential privacy framework for epidemiological modeling with mobile phone data
  • Article
  • Full-text available

October 2023

·

77 Reads

·

5 Citations

·

·

Wanrong Zhang

·

[...]

·

Nishant Kishore

During the COVID-19 pandemic, the use of mobile phone data for monitoring human mobility patterns has become increasingly common, both to study the impact of travel restrictions on population movement and epidemiological modeling. Despite the importance of these data, the use of location information to guide public policy can raise issues of privacy and ethical use. Studies have shown that simple aggregation does not protect the privacy of an individual, and there are no universal standards for aggregation that guarantee anonymity. Newer methods, such as differential privacy, can provide statistically verifiable protection against identifiability but have been largely untested as inputs for compartment models used in infectious disease epidemiology. Our study examines the application of differential privacy as an anonymisation tool in epidemiological models, studying the impact of adding quantifiable statistical noise to mobile phone-based location data on the bias of ten common epidemiological metrics. We find that many epidemiological metrics are preserved and remain close to their non-private values when the true noise state is less than 20, in a count transition matrix, which corresponds to a privacy-less parameter ϵ = 0.05 per release. We show that differential privacy offers a robust approach to preserving individual privacy in mobility data while providing useful population-level insights for public health. Importantly, we have built a modular software pipeline to facilitate the replication and expansion of our framework.

Download

A standardised differential privacy framework for epidemiological modelling with mobile phone data

March 2023

·

45 Reads

During the COVID-19 pandemic, the use of mobile phone data for monitoring human mobility patterns has become increasingly common, both to study the impact of travel restrictions on population movement and epidemiological modelling. Despite the importance of these data, the use of location information to guide public policy can raise issues of privacy and ethical use. Studies have shown that simple aggregation does not protect the privacy of an individual, and there are no universal standards for aggregation that guarantee anonymity. Newer methods, such as differential privacy, can provide statistically verifiable protection against identifiability but have been largely untested as inputs for compartment models used in infectious disease epidemiology. Our study examines the application of differential privacy as an anonymisation tool in epidemiological models, studying the impact of adding quantifiable statistical noise to mobile phone-based location data on the bias of ten common epidemiological metrics. We find that many epidemiological metrics are preserved and remain close to their non-private values when the true noise state is less than 20, in a count transition matrix, which corresponds to a privacy-less parameter ∈ = 0.05 per release. We show that differential privacy offers a robust approach to preserving individual privacy in mobility data while providing useful population-level insights for public health. Importantly, we have built a modular software pipeline to facilitate the replication and expansion of our framework. Author Summary Human mobility data has been used broadly in epidemiological population models to better understand the transmission dynamics of an epidemic, predict its future trajectory, and evaluate potential interventions. The availability and use of these data inherently raises the question of how we can balance individual privacy and the statistical utility of these data. Unfortunately, there are few existing frameworks that allow us to quantify this trade-off. Here, we have developed a framework to implement a differential privacy layer on top of human mobility data which can guarantee a minimum level of privacy protection and evaluate their effects on the statistical utility of model outputs. We show that this set of models and their outputs are resilient to high levels of privacy-preserving noise and suggest a standard privacy threshold with an epsilon of 0.05. Finally, we provide a reproducible framework for public health researchers and data providers to evaluate varying levels of privacy-preserving noise in human mobility data inputs, models, and epidemiological outputs.


FIGURE 1. Weekly COVID-19 case rates among New York City residents 18 to 64 years old, by municipal employment status, July 5 to October 28, 2021. The first and last weeks depicted are partial weeks: week ending July 10 reflects diagnoses during July 5 to 10; week ending October 30 reflects diagnoses during October 24 to 28.
FIGURE 2. Weekly COVID-19 case rates among New York City residents 18 to 64 years old, by municipal employment status, September 23 to November 30, 2021. The first and last weeks depicted are partial weeks: week ending September 25 reflects diagnoses during September 23 to 25; week ending December 4 reflects diagnoses during November 28 to 30.
Effects of Return-to-Office, Public Schools Reopening, and Vaccination Mandates on COVID-19 Cases Among Municipal Employee Residents of New York City

December 2022

·

43 Reads

Journal of occupational and environmental medicine / American College of Occupational and Environmental Medicine

Objective: On September 13, 2021, teleworking ended for New York City municipal employees, and Department of Education (DOE) employees returned to reopened schools. On October 29, COVID-19 vaccination was mandated. We assessed these mandates' short-term effects on disease transmission. Methods: Using difference-in-difference analyses, we calculated COVID-19 incidence rate ratios (IRR) among residents 18-64 years-old by employment status pre- and post-policy implementation. Results: IRRs post- (September 23-October 28) vs. pre- (July 5-September 12) return-to-office were similar between office-based City employees and non-City employees. Among DOE employees, the IRR after schools reopened was elevated 28.4% (95% CI: 17.3%-40.3%). Among City employees, the IRR post- (October 29-November 30) vs. pre- (September 23-October 28) vaccination mandate was lowered 20.1% (95% CI: 13.7%-26.0%). Conclusions: Workforce mandates influenced disease transmission, among other societal effects.


Effects of return-to-office, public schools reopening, and vaccination mandates on COVID-19 cases among municipal employee residents of New York City

October 2022

·

18 Reads

·

1 Citation

Objective: On September 13, 2021, teleworking ended for New York City municipal employees, and Department of Education (DOE) employees returned to reopened schools. On October 29, COVID-19 vaccination was mandated. We assessed these mandates' short-term effects on disease transmission. Methods: Using difference-in-difference analyses, we calculated COVID-19 incidence rate ratios (IRR) among residents 18-64 years-old by employment status pre- and post-policy implementation. Results: IRRs post- (September 23-October 28) vs. pre- (July 5-September 12) return-to-office were similar between office-based City employees and non-City employees. Among DOE employees, the IRR after schools reopened was elevated 28.4% (95% CI: 17.3%-40.3%). Among City employees, the IRR post- (October 29-November 30) vs. pre- (September 23-October 28) vaccination mandate was lowered 20.1% (95% CI: 13.7%-26.0%). Conclusions: Workforce mandates influenced disease transmission, among other societal effects.


Reduced COVID-19 Hospitalizations among New York City Residents Following Age-Based SARS-CoV-2 Vaccine Eligibility: Evidence from a Regression Discontinuity Design

December 2021

·

28 Reads

·

11 Citations

Vaccine X

Background In clinical trials, several SARS-CoV-2 vaccines were shown to reduce risk of severe COVID-19 illness. Local, population-level, real-world evidence of vaccine effectiveness is accumulating. We assessed vaccine effectiveness for community-dwelling New York City (NYC) residents using a quasi-experimental, regression discontinuity design, leveraging a period (January 12–March 9, 2021) when ≥65-year-olds were vaccine-eligible but younger persons, excluding essential workers, were not. Methods We constructed segmented, negative binomial regression models of age-specific COVID-19 hospitalization rates among 45–84-year-old NYC residents during a post-vaccination program implementation period (February 21–April 17, 2021), with a discontinuity at age 65 years. The relationship between age and hospitalization rates in an unvaccinated population was incorporated using a pre-implementation period (December 20, 2020–February 13, 2021). We calculated the rate ratio (RR) and 95% confidence interval (CI) for the interaction between implementation period (pre or post) and age-based eligibility (45–64 or 65–84 years). Analyses were stratified by race/ethnicity and borough of residence. Similar analyses were conducted for COVID-19 deaths. Results Hospitalization rates among 65–84-year-olds decreased from pre- to post-implementation periods (RR 0.85, 95% CI: 0.74–0.97), controlling for trends among 45–64-year-olds. Accordingly, an estimated 721 (95% CI: 126–1,241) hospitalizations were averted. Residents just above the eligibility threshold (65–66-year-olds) had lower hospitalization rates than those below (63–64-year-olds). Racial/ethnic groups and boroughs with higher vaccine coverage generally experienced greater reductions in RR point estimates. Uncertainty was greater for the decrease in COVID-19 death rates (RR 0.85, 95% CI: 0.66–1.10). Conclusion The vaccination program in NYC reduced COVID-19 hospitalizations among the initially age-eligible ≥65-year-old population by approximately 15% in the first eight weeks. The real-world evidence of vaccine effectiveness makes it more imperative to improve vaccine access and uptake to reduce inequities in COVID-19 outcomes.


Figure 1: Counties included in the study by their NCHS category (A) and the number of observations of counties with at least 100 cases with available estimates, by state and epidemiological week (B) NCHS=National Center for Health Statistics.
Figure 2: Example of metrics used for analysis for Shelby County, Tennessee, which is categorised as a large central metro by the NCHS with a LOESS curve added NCHS=National Center for Health Statistics. LOESS=locally estimated scatterplot smoothing. % change in=percentage change of movement into counties. % change out=percentage change of movement out of counties. % change within=percentage change of movement within counties. Dwell=average number of locations users visited. Entropy=average entropy of movement. R t =effective reproduction number.
Figure 5: Correlation between log (R t ) and the modelled log (R t ) over the course of the study period R t =effective reproduction number.
Evaluating the reliability of mobility metrics from aggregated mobile phone data as proxies for SARS-CoV-2 transmission in the USA: a population-based study

November 2021

·

77 Reads

·

34 Citations

The Lancet Digital Health

Background: In early 2020, the response to the SARS-CoV-2 pandemic focused on non-pharmaceutical interventions, some of which aimed to reduce transmission by changing mixing patterns between people. Aggregated location data from mobile phones are an important source of real-time information about human mobility on a population level, but the degree to which these mobility metrics capture the relevant contact patterns of individuals at risk of transmitting SARS-CoV-2 is not clear. In this study we describe changes in the relationship between mobile phone data and SARS-CoV-2 transmission in the USA. Methods: In this population-based study, we collected epidemiological data on COVID-19 cases and deaths, as well as human mobility metrics collated by advertisement technology that was derived from global positioning systems, from 1396 counties across the USA that had at least 100 laboratory-confirmed cases of COVID-19. We grouped these counties into six ordinal categories, defined by the National Center for Health Statistics (NCHS) and graded from urban to rural, and quantified the changes in COVID-19 transmission using estimates of the effective reproduction number (Rt) between Jan 22 and July 9, 2020, to investigate the relationship between aggregated mobility metrics and epidemic trajectory. For each county, we model the time series of Rt values with mobility proxies. Findings: We show that the reproduction number is most strongly associated with mobility proxies for change in the travel into counties (0·757 [95% CI 0·689 to 0·857]), but this relationship primarily holds for counties in the three most urban categories as defined by the NCHS. This relationship weakens considerably after the initial 15 weeks of the epidemic (0·442 [-0·492 to -0·392]), consistent with the emergence of more complex local policies and behaviours, including masking. Interpretation: Our study shows that the integration of mobility metrics into retrospective modelling efforts can be useful in identifying links between these metrics and Rt. Importantly, we highlight potential issues in the data generation process for transmission indicators derived from mobile phone data, representativeness, and equity of access, which must be addressed to improve the interpretability of these data in public health. Funding: There was no funding source for this study.


Reduced COVID-19 Hospitalizations among New York City Residents Following Age-Based SARS-CoV-2 Vaccine Eligibility: Evidence from a Regression Discontinuity Design

July 2021

·

48 Reads

Background: In clinical trials, several SARS-CoV-2 vaccines were shown to reduce risk of severe COVID-19 illness. Local, population-level, real-world evidence of vaccine effectiveness is accumulating. We assessed vaccine effectiveness for community-dwelling New York City (NYC) residents using a quasi-experimental, regression discontinuity design, leveraging a period (January 12-March 9, 2021) when ≥65-year-olds were vaccine-eligible but younger persons, excluding essential workers, were not. Methods: We constructed segmented, negative binomial regression models of age-specific COVID-19 hospitalization rates among 45-84-year-old NYC residents during a post-vaccination program implementation period (February 21-April 17, 2021), with a discontinuity at age 65 years. The relationship between age and hospitalization rates in an unvaccinated population was incorporated using a pre-implementation period (December 20, 2020-February 13, 2021). We calculated the rate ratio (RR) and 95% confidence interval (CI) for the interaction between implementation period (pre or post) and age-based eligibility (45-64 or 65-84 years). Analyses were stratified by race/ethnicity and borough of residence. Similar analyses were conducted for COVID-19 deaths. Results: Hospitalization rates among 65-84-year-olds decreased from pre- to post-implementation periods (RR 0.85, 95% CI: 0.74-0.97), controlling for trends among 45-64-year-olds. Accordingly, an estimated 721 (95% CI: 126-1,241) hospitalizations were averted. Residents just above the eligibility threshold (65-66-year-olds) had lower hospitalization rates than those below (63-64-year-olds). Racial/ethnic groups and boroughs with higher vaccine coverage generally experienced greater reductions in RR point estimates. Uncertainty was greater for the decrease in COVID-19 death rates (RR 0.85, 95% CI: 0.66-1.10). Conclusion: The vaccination program in NYC reduced COVID-19 hospitalizations among the initially age-eligible ≥65-year-old population by approximately 15%. The real-world evidence of vaccine effectiveness makes it more imperative to improve vaccine access and uptake to reduce inequities in COVID-19 outcomes.


Figure 6: Correlation between log(R(t)) and the predicted log(R(t)) over the course of the study period. The model results are highly correlated with R(t) in the first half of the study period but performance drops off in the second half.
The relationship between human mobility measures and SAR-Cov-2 transmission varies by epidemic phase and urbanicity: results from the United States

April 2021

·

46 Reads

·

2 Citations

Global efforts to prevent the spread of the SARS-COV-2 pandemic in early 2020 focused on non-pharmaceutical interventions like social distancing; policies that aim to reduce transmission by changing mixing patterns between people. As countries have implemented these interventions, aggregated location data from mobile phones have become an important source of real-time information about human mobility and behavioral changes on a population level. Human activity measured using mobile phones reflects the aggregate behavior of a subset of people, and although metrics of mobility are related to contact patterns between people that spread the coronavirus, they do not provide a direct measure. In this study, we use results from a nowcasting approach from 1,396 counties across the US between January 22nd, 2020 and July 9th, 2020 to determine the effective reproductive number (R(t)) along an urban/rural gradient. For each county, we compare the time series of R(t) values with mobility proxies from mobile phone data from Camber Systems, an aggregator of mobility data from various providers in the United States. We show that the reproduction number is most strongly associated with mobility proxies for change in the travel into counties compared to baseline, but that the relationship weakens considerably after the initial 15 weeks of the epidemic, consistent with the emergence of a more complex ecosystem of local policies and behaviors including masking. Importantly, we highlight potential issues in the data generation process, representativeness and equity of access which must be addressed to allow for general use of these data in public health.


Lockdowns result in changes in human mobility which may impact the epidemiologic dynamics of SARS-CoV-2

March 2021

·

94 Reads

·

56 Citations

In response to the SARS-CoV-2 pandemic, unprecedented travel restrictions and stay-at-home orders were enacted around the world. Ultimately, the public’s response to announcements of lockdowns—defined as restrictions on both local movement or long distance travel—will determine how effective these kinds of interventions are. Here, we evaluate the effects of lockdowns on human mobility and simulate how these changes may affect epidemic spread by analyzing aggregated mobility data from mobile phones. We show that in 2020 following lockdown announcements but prior to their implementation, both local and long distance movement increased in multiple locations, and urban-to-rural migration was observed around the world. To examine how these behavioral responses to lockdown policies may contribute to epidemic spread, we developed a simple agent-based spatial model. Our model shows that this increased movement has the potential to increase seeding of the epidemic in less urban areas, which could undermine the goal of the lockdown in preventing disease spread. Lockdowns play a key role in reducing contacts and controlling outbreaks, but appropriate messaging surrounding their announcement and careful evaluation of changes in mobility are needed to mitigate the possible unintended consequences.


Cell-phone traces reveal infection-associated behavioral change

February 2021

·

84 Reads

·

13 Citations

Proceedings of the National Academy of Sciences

Significance Infectious disease control critically depends on surveillance and predictive modeling of outbreaks. We argue that routine mobile-phone use can provide a source of infectious disease information via the measurements of behavioral changes in call-detail records (CDRs) collected for billing. In anonymous CDR metadata linked with individual health information from the A(H1N1)pdm09 outbreak in Iceland, we observe that people moved significantly less and placed fewer, but longer, calls in the few days around diagnosis than normal. These results suggest that disease-transmission models should explicitly consider behavior changes during outbreaks and advance mobile-phone traces as a potential universal data source for such efforts.


Citations (17)


... While our analysis uses these data to highlight disparities, it is essential to emphasize the importance of cautious data handling by researchers to prevent misuse, such as racial profiling. Fortunately, current methods including data aggregation and differential privacy could help mitigate these risks, enabling meaningful analysis while upholding ethical standards (De Montjoye et al 2018, Savi et al 2023). ...

Reference:

Uncovering disparities in water-based outdoor recreation using cell phone mobility data
A standardised differential privacy framework for epidemiological modeling with mobile phone data

... Consequently, we applied the RDD using birth month as the running variable with a cutoff of March 1957. Several studies have estimated the causal relationships between influenza, COVID-19 vaccinations, and health outcomes using RDD with age or birth timing as the running variables [18][19][20][21]. However, to the best of our knowledge, few studies have investigated the causal relationship between COVID-19 vaccination and protective behaviors against COVID-19. ...

Reduced COVID-19 Hospitalizations among New York City Residents Following Age-Based SARS-CoV-2 Vaccine Eligibility: Evidence from a Regression Discontinuity Design

Vaccine X

... Furthermore, translating mobility changes into contact reductions remains an open challenge. Therefore, the performance of the data-driven model may vary depending on different approaches to effective contacts rescaling [74][75][76][77][78]. Finally, our forecasts are based on data reported as of today, not addressing the challenge of retrospective data adjustments (i.e., backfilling) very common for epidemiological datasets. ...

Evaluating the reliability of mobility metrics from aggregated mobile phone data as proxies for SARS-CoV-2 transmission in the USA: a population-based study

The Lancet Digital Health

... While mobility metrics can potentially behave as proxies for epidemiological measures, such as contact rate and effective reproductive number, the intensity and direction of this relationship can change by epidemic stage 9 . For example, early in the pandemic, before mask mandates were widely adopted, a measure of the proportion of individuals in a county who spent time outside their home was a useful proxy for potentially contagious contacts. ...

The relationship between human mobility measures and SAR-Cov-2 transmission varies by epidemic phase and urbanicity: results from the United States

... While much research has been conducted on the economic and social consequences of the epidemic, the interaction between climate conditions and government policies remains unexplored. Understanding how environment impacts the spread of respiratory viruses such as COVID-19 can help public health officials respond more effectively and prepare for future pandemics [2][3][4]. ...

Lockdowns result in changes in human mobility which may impact the epidemiologic dynamics of SARS-CoV-2

... 25 An analysis of cell-phone call records during the 2009 H1N1v pandemic found reduced mobility among individuals diagnosed with influenza-like illness. 26 For COVID-19, studies have identified variations in heart rate, HRV, steps, and respiratory rate preceding symptom onset, 18,27,28 emphasizing the importance of recognizing early markers for timely interventions. Yet, a holistic understanding of behavioral and physiological shifts during the diagnostic period remains incomplete. ...

Cell-phone traces reveal infection-associated behavioral change

Proceedings of the National Academy of Sciences

... (Lough, 2022). This interconnectedness of responders and organizations reduces redundancy in efforts and leads to a more unified approach to complex humanitarian situations (Acosta et al., 2020). ...

Quantifying the dynamics of migration after Hurricane Maria in Puerto Rico

Proceedings of the National Academy of Sciences

... Aggregating properly mobile phone traces means to assess how many details of them are needed to describe epidemiological links between locations, and which details are instead negligible, or noise. Considering the privacy issues on dealing with mobile phone data and the time needed to extract and preprocessed them before their sharing with researchers [30][31][32][33][34] , to know a priori the resolution needed to inform an epidemic model become crucial in the management of a crisis. ...

The use of mobile phone data to inform analysis of COVID-19 pandemic epidemiology

... The recorded changes in the contact patterns are concerning in the contexts of epidemics of respiratory pathogens such as influenza and SARS-CoV-2. The rate of local contacts was known to be positively associated with the local reproduction ratio in the context of the COVID-19 pandemic [45,46]. As such, the evidence of rising household and community contacts in the aftermath of floods may increase the chance of outbreaks by temporarily lowering the epidemic threshold [47], or facilitate epidemic circulation of ongoing outbreaks. ...

Reductions in commuting mobility correlate with geographic differences in SARS-CoV-2 prevalence in New York City

... In most countries, mobile phone billing is proportional to the number of call and SMS, generating a large inequality in individuals' activity and thus in data representativeness between social strata 42 . Data is indeed most representative of urban populations for target subgroups, suffering from demographic and geographical representativeness, and other issues concerning ownership, mobile phones sharing and, heterogeneity in cell tower distribution and in individuals' [42][43][44][45] . However, the aggregation of mobile phone individual trajectories into coupling forces thought a metapopulation approach allows to partially avoid mobile phone biases [46][47][48] . ...

Measuring mobility to monitor travel and physical distancing interventions: a common framework for mobile phone data analysis

The Lancet Digital Health