Search results

Comparative effectiveness research compares two or more treatments to see which one works best for which patients. But patient traits, such as age or income, may affect patients' treatment choices. These traits may also affect patients' responses to treatments. As a result, researchers may have trouble telling whether a patient's traits, the treatment, or a mix of the two affected how well a treatment worked.

Statistical methods called matching methods can help address this problem when researchers use patient data to compare the effects of treatments. Matching methods help researchers find data from patients who had similar traits such as age or race and received different treatments. Because the patients are similar except for the treatment they receive, the differences in patients' health can more likely be credited to the treatment. Existing methods work well for comparing up to two treatments. But they may not work with three or more treatments.

In this study, the research team created two new matching methods to compare the effects of three or more treatments. The team then analyzed the new methods under different conditions to see how well each worked."

A randomized trial is one of the best ways to learn if one treatment works better than another. Randomized trials assign patients to different treatments by chance. But they are not always affordable, and they take a long time to complete.

When randomized trials aren't possible, researchers can use observational studies to learn how treatments work. In observational studies, researchers look at what happens when patients and their doctors choose the treatments. Traits such as age or health may affect treatment choices. These traits may also affect patients' responses to treatment, making it hard to know if the treatment or the traits affected the patients' responses.

Some study designs and statistical methods may help address this problem and make results from observational studies more useful. These methods can give researchers more data about whether treatments work and how the same treatment can affect groups of patients differently.

The research team conducted three studies to test different methods of designing and analyzing observational studies. They wanted to know if observational studies that used these methods produced results similar to randomized trials.

Propensity scores (PS) and instrumental variables (IV) are methods used to assess treatment effects in observational studies when randomized controlled trials (RCTs) are not feasible. However, these methods have limitations, especially when using high-dimensional data, or data with numerous variables or many non-linear and interaction terms. Choices on which variables and non-linear and interaction terms to include may lead to model misspecification. The objective of this study was to develop and test a set of PS and IV methods that account for model misspecification when estimating causal effects of treatments using high-dimensional data.

First, the research team created the two new methods for use with high-dimensional data. The team then used a computer program to create test data that look like real patient data. The team applied the new methods to the test data. Next, the research team applied the new methods to real data from previous studies. They applied the PS method to data from Connors et al. (1996) and applied the IV method to data used by Card (1995). Using both test and real data, the research team compared findings from the new methods with those from existing PS and IV methods and checked to see if findings from the new methods were accurate when including different patient traits and health conditions in the analysis.

This collection contains the R software package RCAL and accompanying documentation. The package source as a .tar.gz file and six different versions are available in a zipped package. Files have been released as received by ICPSR from the depositor:

• For R version 4.2, created April 24, 2022 (Windows, r-oldrel)
• For R version 4.3, created October 20, 2023 (Windows, r-release)
• For R version 4.4, created March 14, 2024 (Windows, r-devel)
• For R version 4.2, created April 1, 2023 (Mac, arm64, r-oldrel)
• For R version 4.3, created April 6, 2023 (Mac, arm64, r-release)
• For R version 4.3, created April 11, 2023 (Mac, x86_64, r-release)

Comparative effectiveness research compares two or more treatments to see which one works better for which patients. In some studies, researchers assign patients by chance to several treatments or to have or not have a treatment. But approaches that assign patients by chance are not always suitable. For example, assigning patients to a new treatment may not be good medical care.

For this reason, researchers sometimes do studies using data collected when patients and their doctors choose the treatments. Data from such studies are observational data. When using observational data for research, it can be hard to know if the effect of a treatment is because of the treatment or other factors, such as patients' age or health history. In these cases, researchers use statistical methods to understand the effect of the treatment. Depending on the study's focus and design, some methods work better than others.

In this study, the research team developed guidance for researchers to help them choose methods for their study.

To access the methods and software, please visit the DECODE CER Tool website.

To study the impact of COVID-19 pandemic on frontline healthcare workers in the United States over time, the Healthcare Worker Exposure Response and Outcomes (HERO) Registry was created in 2020 to form a virtual research community of healthcare workers (and later, their family members and community members). The registry was intended for healthcare workers interested in completing research studies related to the COVID-19 pandemic and its impacts on their lives. Observational data were collected at various timepoints between April 2020 and September 2022 via web-based questionnaires available on the HERO Registry online portal.

This collection contains 39 sets of data from over 50,000 HERO Registry members. Datasets represent separate surveys with distinct survey designs and sampling criteria. Surveys focused on health history, workplace experiences, COVID-19 exposure, social support, mental health, and the respondents' willingness to remain in or leave the healthcare field. Datasets 24 through 39 represent "hot topics" such as vaccines, vaccine willingness and uptake, childcare and school arrangements, and staffing shortages. Datasets for registry administration, respondent demographics, and survey eligibility criteria are also included.

In the United States, tobacco smoking is associated with significant morbidity and premature mortality for individuals with serious mental illness (SMI) (e.g., schizophrenia, post-traumatic stress disorder, bipolar disorder, major depressive disorder). While many smokers with SMI wish to quit smoking, few are offered advice or treatments with demonstrated effectiveness in reducing tobacco dependence, primarily medication-assisted treatments. The overall aim of this randomized controlled trial was to test the effects of provider education (PE) (i.e. provider-level educational intervention focused on evidence-based smoking cessation treatment for those with SMI) and community health worker (CHW) support on the provision and utilization of smoking cessation treatment to those with SMI, and cessation rates for adults with SMI who smoke or use tobacco over a 2-year period. The objectives of this trial were to:

1. Examine whether an intervention combining PE and CHW support would increase prescriber provision of advice and assistance to quit smoking, and improve tobacco cessation rates in smokers with SMI compared to usual care/treatment as usual (TAU) and compared to PE-only treatment
2. Determine the effect of the combined PE+CHW intervention on patient-reported overall health compared to TAU and PE-only treatment

Eligible individuals were recruited from two outpatient psychiatric service providers in the Boston, Massachusetts metropolitan area. Clinics where individuals received services were randomized into either the TAU condition or into the PE condition, where health care providers would receive additional education on first-line medications used to treat tobacco use disorder. Within clinics in the PE arm, individuals were further randomized into the community health worker (CHW) support condition (PE+CHW), where CHWs would assist participants with smoking cessation care access and provide community outreach and education, or no CHW support (PE-only). Enrolled participants (n=1,010) completed surveys on smoking/tobacco use at 3 timepoints: study baseline, 1 year post-randomization, and 2 years post-randomization.

A mixed-methods evaluation of the trial was also conducted post-intervention, using an interactive convergent design. The aims of the evaluation were to identify barriers and facilitators to effective implementation; examine how primary care providers differed by performance and engagement level, and how experiences with the intervention compared across these groups; and identify anticipated barriers to implementing the intervention as discussed by stakeholders. Quantitative outcome and visit data from the trial were used in the evaluation. For the evaluation's qualitative component, interviews were conducted with purposively sampled community health workers, smoker participants, primary care providers, and other stakeholders in policy, payor, and clinical administration. Please note that the qualitative evaluation data are not available for this collection.

For people with serious chronic conditions, healthcare that defaults to all available treatments without considering patient preferences risks harms that may exceed benefits. Advance care planning (ACP) has the potential to align healthcare with what is important to patients and maximize quality of life. While primary care is where most people receive most of their care, engaging patients in ACP is not routine in primary care given competing demands and limited resources. Primary care clinicians, patients, and families agree that it is preferred to make plans before there is a medical crisis. The research team's goal was to make ACP routine in primary care and to "move it upstream" so that it included improving the quality of the last years of life as well as respecting wishes for end of life care.

This study included a comparative effectiveness trial of team-based versus individual clinician-focused ACP in primary care practices. The research team adapted Ariadne Labs' Serious Illness Care Program (SICP) and aimed to determine if a team approach produces better patient outcomes and explore factors influencing implementation of ACP across practices.

Seven practice-based research networks (PBRNs) in the United States and Canada randomized their primary care practices to team-based or individual clinician-focused versions of SICP. Team members and clinicians completed training, and implementation was supported through practice facilitation. Consented patient participants completed a baseline survey after initial conversations and follow-up surveys at 6 and 12 months later. Forty practices (21 team, 19 clinician) completed training and referred patients to the study. Half of the practices were rural, 80 percent were family medicine, and 33 percent were medical residency training sites. 535 healthcare staff completed training. Both arms trained primary care providers; the team arm also trained nurses, medical assistants, and other roles. 1,321 patients and care partners were referred; and 917 consented and were enrolled (455 from team practices, 462 from clinician). Data from 802 patients were included in the primary analyses. Qualitative implementation data was collected during practice facilitation and from practice interviews.

This collection includes quantitative data collected from primary care practices (DS1) and team members and clinicians (DS2) from study sites located in the United States.

Using a pragmatic trial design to limit exclusions, the investigators conducted a 36-month multi-center randomized effectiveness trial to compare the impact of an enhanced usual care (control) intervention, with exercise coaching (exercise), on fragility fractures and serious fall-related injuries (FF/SFRI) in patients with a previous fragility fracture. Specifically, the investigators examined the impact of the intervention on social loneliness, physical function, and bone strength. 1,139 individuals over 65 with a history of fragility fractures and/or osteoporosis were recruited over two years across three regions of Pennsylvania and randomized into either the enhanced usual care control group or exercise with coaching treatment group, where in-person exercise activities were led by trained volunteers.

Dataset (DS) 1 contains the following data used for analysis: participant characteristics at baseline by study group (referred to as Table 5 in the documentation), intervention participant characteristics at baseline based on exercise session type (referred to as Table 6), cumulative incidences of first serious fall-related injury compared by study group (referred to as Figure 3), cumulative incidence for first serious fall-related injury by age, gender, race, and osteoporosis medication (referred to as Table 8 and Figure 4), and cumulative incidence for first series fall-related injury by tertile of average intervention sessions per month (referred to as Figure 5). Other datasets used for analysis are fall injury data (DS2), monthly workout sessions data (DS3), secondary outcomes data (DS4, referred to as Table 7), and adverse events data (DS5, referred to as Table 9). DS6 includes markers designating before and after the start of the COVID-19 pandemic (March 11, 2020), allowing for analyses of participants who experienced fall-related injuries relative to COVID-19.

Datasets labeled "Miscellaneous" were not used in any analysis. These datasets contain extra measures from screening (DS7), baseline assessments (DS8), 4-month check-in visits (DS9), participant's distance to study site (DS10), coaching check-ins for weeks 1-12 (DS11), exercise sessions by month (DS12), adverse events (DS13), and end of study information (DS14).

The Preserving Kidney Function in Children With Chronic Kidney Disease (PRESERVE) study was designed to provide new knowledge to inform shared decision-making regarding blood pressure (BP) management for pediatric chronic kidney disease (CKD). PRESERVE compared the effectiveness of alternative strategies for monitoring and treating hypertension on preserving kidney function; expanded the National Patient-Centered Clinical Research Network (PCORnet) Common Data Model by adding pediatric- and kidney-specific variables and linking electronic health record data to other kidney disease databases; and assessed the lived experiences of patients related to BP management.

Participants were recruited from 15 clinical institutions across the United States. The research team analyzed electronic health record (EHR) data from 11,851 children with CKD and their caregivers to compare different ways to monitor and treat BP to preserve kidney function. In addition, a subset of patients and caregivers completed an online survey detailing patient-reported outcomes, such as fatigue, life satisfaction, pain levels, sleep disturbance, anxiety, and peer relationships (n=395).

Due to the risk of re-identification based on unique patterns in the individual-level PCORnet electronic health record (EHR) data, patient privacy regulations prohibit the public release of the individual-level data. This collection contains the code underlying the analysis; instructions, codesets, and output lists for the PCORnet queries; and the survey questionnaires for patients and family members.

Comparative effectiveness research compares two or more treatments to see which one works better for which patients. When researchers can't assign patients by chance to treatments, they can use observational studies. In observational studies, researchers use data like health records to compare treatment effects. But it can be hard to know if the effects are due to the treatment or to patient traits, like age.

To address this issue, researchers can use statistical methods called propensity score matching, or PSM. With PSM, researchers create groups of patients for analysis who have received different treatments. They match patients with similar traits across groups. This method reduces bias when comparing treatments. But current PSM methods don't work well or may take many hours when comparing three or more treatments or when using large data sets.

In this study, the research team created and tested a new method for matching patients from large data sets to compare the effects of three or more treatments.

The COVID-19 Citizen Science (CCS) Study was launched early in the pandemic to collect patient-reported information about exposures, risk behaviors and outcomes relevant to the pandemic. The Patient-Centered Outcomes Research Institute (PCORI) funded the research team to expand recruitment into CCS using PCORnet, the National Patient-Centered Clinical Research Network, and to use the resulting data to compare the patient-reported impact of pandemic associated policies. The research team systematically collected pandemic-associated policies enacted by counties across the United States (focusing in areas where there were many CCS participants), and to do so on a weekly basis from the beginning of the pandemic using publicly available sources.

Researchers combined data from various sources to answer two primary research questions (RQ):

1. What is the comparative impact of different shelter-in-place/reopening policies, overall and in vulnerable populations, on patient-reported financial insecurity, mental health, and other subjective outcomes important to patients?
2. What is the comparative effectiveness of county-level containment and mitigation strategies at achieving timely access to COVID-19 vaccination, testing, healthcare, information and contact tracing?

The research team collected patient-reported data from the CCS study and policy data from the U.S COVID-19 County Policy (UCCP) database. Electronic health record (EHR) data were also available from some participants recruited from health systems located across 7 U.S. states who consented and authorized use of these data for the study. Data for these participants were extracted from the PCORnet Common Data Model (CDM). Additional county-level contextual variables were included in analysis.

This collection contains CCS survey data on patient-reported anxiety with county-level policies data (DS1), respondent demographics (DS2), baseline survey results (DS3), daily (DS4) and weekly (DS5) COVID-19 symptoms reports, COVID-19 vaccination surveys repeated monthly (DS6) as well as a one-time vaccination survey (DS7), and pandemic impacts check-in surveys (DS8). CDM datasets include logistic regression model outcomes to predict study enrollment among all invited participants (DS9), codes for immunizations (DS10), laboratory tests (DS11), and procedures (DS12). County-level variables are also available for years 2021 (DS13) and 2023 (DS14).

Comparative effectiveness research compares two or more treatments to see which one works better for certain patients. For example, research can see if medicines or stents work better for people with heart problems. Such research may include:

• Observational studies. A research team studies what happens when patients and their clinicians choose the treatments. Traits, such as age or health, may affect patients' treatment choices. These traits may also affect patients' responses to treatments. It may be hard for the team to tell if a patient's traits, the treatment, or a mix of the two affected how well the treatment worked.
• Clinical trials. The team assigns patients to a treatment by chance. Traits may affect a patient's ability to join a clinical trial.

One goal of comparative effectiveness research is to find out which treatments work best for different groups of patients. For example, treatments may work differently for patients with only one health problem than for those with more than one health problem.

In observational studies, researchers look at health outcomes when patients and their doctors choose the treatments. These studies often use data from electronic health records, or EHRs. Researchers can apply propensity score, or PS, methods to look at different groups of patients. With PS methods, researchers create groups of patients with similar traits who had different treatments. But PS methods require researchers to have data on all patient traits that could affect how well the treatment works. With EHR data, data on some patient traits, like health problems, may be missing. Using current PS methods in observational studies may lead to biased results.

In this study, the research team created new guidance for using PS methods with EHR data to look at the effects of treatment in different groups of patients. The team also created and tested new PS methods to make groups of patients with similar traits.

The overall goal for this project was to reduce the incidence of COVID-19, hospitalization, and mortality among adults with serious mental illness (SMI) and intellectual disabilities/developmental disabilities (IDD) in congregate living settings (i.e., group homes) in Massachusetts, as well as to reduce COVID-19 incidence among staff who work in these settings. The research team was guided by two comparative effectiveness questions:

1. With the goal of prioritizing and making actionable best practices available as resources, what is the comparative effectiveness of various types and intensities of preventative interventions (e.g., screening, isolation, contact tracing, hand hygiene, physical distancing, use of face masks) in reducing rates of COVID-19, related hospitalizations, and related mortality in this population?
2. With the goal of effectively implementing best practices, what is the most effective implementation strategy to reduce rates of COVID-19 in this population: using tailored best practices (TBP) with SMI/IDD residents and staff of group homes in mind, or general best practices (GBP) from state and federal standard guidelines for all congregate care settings?

The specific aims of this study were as follows:

Aim 1a. Synthesize existing baseline data collected by 6 state behavioral health agencies on COVID-19 rates, hospitalization, mortality, and use of infection prevention practices.

Aim 1b. Collect stakeholder input via surveys and virtual focus groups on staff and resident experiences and on barriers/facilitators to implementing recommended preventative practices.

Aims 2a and 2b. Determine the comparative effectiveness of various COVID-19 preventative practices by (Aim 2a) using a validated simulation model to estimate COVID-19 spread in group homes and (Aim 2b) obtaining stakeholder input on prioritizing and defining tailored best practices for implementation.

Aim 3. Compare the effectiveness of TBPs with GBPs by using a hybrid effectiveness-implementation cluster randomized controlled trial.

Data collected to answer Aims 1 and 2 served as the foundation for designing the Aim 3 trial. Data for the trial were collected in 3-month intervals beginning January 2021 (baseline) until October 2022 (15-month follow-up). Residents and staff were sampled from approximately 400 group homes. Primary implementation outcome measures were COVID-19 vaccination rates and fidelity scores. The primary effectiveness outcome measure was COVID-19 infection.

Notes: This collection contains only data from Aim 1a and Aim 3. Throughout the data and documentation, "intellectual and/or developmental disabilities" is abbreviated as both IDD and ID/DD.

Comparative effectiveness research compares two or more treatments to see which one works better for which patients. One type of research study is a randomized controlled trial, or an RCT. In an RCT, the research team assigns patients to a treatment by chance.

Other types of studies use information from health records and registries. Registries store data about patients with a specific health problem. They often include information on how each patient responds to a treatment. Because researchers don't assign treatments by chance in such studies, differences in how patients respond to a treatment may be from the treatment or something else, such as a patient's age or the severity of their illness. In studies using registries and health records, researchers apply statistical approaches, called causal inference methods, to estimate how treatments work. At the same time, they look at other things that could affect results, like a patient's age.

Researchers can choose among many different causal inference methods. But they may have a hard time knowing which methods to use or how to use complex methods correctly. In this study, the research team made an interactive online guide for researchers. The guide, called CERBOT, helps researchers design studies and select these methods.

A randomized controlled trial, or RCT, is often the best way to learn if one treatment works better than another. RCTs assign patients to different treatments by chance. But RCTs are not always feasible. In such cases, researchers can use observational studies. In observational studies, researchers look at what happens when patients and their doctors choose the treatments. Traits such as age, gender, or health status may affect treatment choices. These traits may also affect patients' health, making it hard to know if changes in patients' health are due to treatment or to patient traits.

To figure out whether changes in patients' health result from treatment or something else, researchers use statistical methods. Two of these methods are:

• Propensity score, or PS. PS methods compare the health of patients who have similar measured traits but received different treatments. These traits are in patient health records.
• Instrumental variable, or IV. IV methods account for things that may affect treatment choice and patients' health but aren't in the patients' health records, such as personal preference about treatment.

But existing PS and IV methods don't work well when data sets include a lot of traits and health conditions for each patient. Such data sets are called high-dimensional data. In this study, the research team created and tested one PS method and one IV method for use with high-dimensional data.

Patients may respond differently to the same treatment due to individual traits such as age or gender. Knowing how different traits can affect a patient's response to treatment can help doctors and patients make better treatment decisions. For example, this information can help doctors know what types of cancer medicines work better for certain patients. This project focuses on improving the methods that researchers use to compare how treatments work for different patients.

In this project, the research team developed and tested a statistical method called random forests, or RF. RF is a way to analyze data using a technique called machine learning. In machine learning, computers use data to learn how to perform different tasks with little or no human input. Many types of RF methods exist. The team compared multiple RF methods to learn how well the methods would work to find out how patients with different traits respond to the same treatment.

To access the R package, please visit the randomForestSRC CRAN webpage.

Treatment plans for patients with long-term health problems such as diabetes or arthritis often change over time. Such plans are called adaptive treatment plans as doctors adapt treatment based on the patient's health problem and response to earlier treatments. Adaptive treatment plans are common, but the methods to assess how well a plan works may not always provide accurate results. To know which plans are best for patients, researchers need better methods to compare these adaptive plans.

In this study, the research team developed and tested a new statistical method and looked at whether it could more accurately compare adaptive treatment plans.

To access the methods and software, please visit the PCATS Application.

Patients may respond differently to the same treatment due to differences in personal traits such as age, gender, or the number and type of health problems they have. Researchers use statistical methods to predict how well a treatment may work for patients based on their personal traits. But current methods may not work well if patients have many health problems or are taking other medicines.

In this project, the research team created new methods to figure out which patient traits are related to treatment benefits to help doctors and patients understand the likely treatment benefits for individual patients.

To access the methods, software, and R package, please visit the personalized CRAN webpage and personalized GitHub

The TARGET (Targeted Interventions to Prevent Chronic Low Back Pain in High-Risk Patients) Trial was a primary care-based, multisite, cluster randomized, pragmatic trial comparing guideline-based care (GBC) to GBC + referral to Psychologically Informed Physical Therapy (PIPT) for patients presenting with acute lower back pain (LBP) and identified as high risk for persistent disabling symptoms. Chronic lower back pain (LBP) is defined as a response of "more than three months" to question 1, and a response of "half the days or more than half the days" in the past 6 months to question 2. See Appendix 1 for the LBP Questionnaire in the Protocol report.

Study sites included primary care clinics within each of four geographical regions in the United States, with clinics randomized to either GBC or GBC+PIPT. Acute LBP patients at all clinics were risk stratified (high, medium, low) using the STarT Back Tool. The primary outcomes were the presence of chronic LBP and LBP-related functional disability determined by the Oswestry Disability Index at 6 months. Secondary outcomes were LBP-related processes of health care and utilization of services over 12 months, determined through electronic medical records.

Study enrollment began in May 2016 and concluded in June 2018. The trial was powered to include at least 1,860 high-risk patients in the cluster-randomized controlled trial cohort. A prospective observational cohort of approximately 6,900 low and medium-risk acute LBP patients was enrolled concurrently.

This data collection contains a single data file with 223 variables and 9,730 cases. The number of respondents at each of the study locations were:

• Boston Medical Center: 997 respondents
• Intermountain Health (Salt Lake City): 2,094 respondents
• Johns Hopkins University (Baltimore): 1,615 respondents
• University of Pittsburg Medical Center: 5,024 respondents

Researchers can use data from electronic health records, or EHRs, in studies that compare two or more treatments. In these studies, researchers need to identify all patients with the same phenotype. Phenotypes are a person's known traits, like height and weight, or known health problems, like diabetes. However, in EHR data, some data on patient traits or health problems may be missing for some patients.

Missing data in EHRs make it hard to correctly identify all patients with the same phenotype. It's even harder when data are missing due to a patient's health status. For example, patients with uncontrolled diabetes may need more lab tests than patients with controlled diabetes. As a result, researchers who are looking at lab tests may not identify patients with controlled diabetes as having diabetes.

In this project, the research team developed and tested a new statistical method that accounts for missing EHR data to estimate patient phenotypes.

To access the methods and software, please visit the bias_correction GitHub repository.

Data collected from interviews and group discussions, called qualitative data, can help researchers understand people's experiences, values, and cultures. But large amounts of qualitative data can be hard to show in a way that's easy for people to understand.

In this study, the research team created charts called ethnoarrays. These charts use color coding to show individual stories and overall patterns in qualitative data. The team wanted to learn whether ethnoarrays were useful and easy to understand.

Comparative effectiveness research compares two or more treatments to see which treatment works better for which patients. Such research may include

• Randomized controlled trials, or RCTs. Researchers assign patients to a treatment by chance. Researchers consider RCTs to be the best way to figure out when changes in patients' health result from the treatment.
• Observational studies. Researchers study what happens when patients and their doctors choose treatments. Patient traits, such as age or health, may affect treatment choices. These traits may also affect patients' responses to treatments. Determining whether a patient's traits, the treatment, or a mix of the two affected how well the treatment worked may be difficult.

In observational studies, researchers use statistical methods to help find out whether changes in patients' health result from treatment or something else. Existing methods work well when studies look at whether treatment affects the risk of a health event, such as a heart attack. In these cases, researchers can compare how often patients had heart attacks before and after patients receive treatment. But existing methods don't work well when studies look at the risk of a one-time event, such as death.

In this study, the research team tested a new statistical method for observational studies called posttreatment event rate ratio, or PTERR, that helps figure out whether a treatment reduces the risk of death.

Randomized controlled trials, or RCTs, look at how well treatments work. But people who take part in RCTs may differ from patients who receive care in clinics. For instance, patients who take part in RCTs may be less likely to smoke or may have fewer health problems. These differences can affect how well a treatment works. As a result, a treatment may work differently for a patient receiving care in a clinic than it did for patients who took part in the RCT.

Researchers can use statistical methods to account for differences in patient traits and behaviors. In this project, the research team developed and tested new methods to account for these differences. They used the methods to apply RCT results to patients receiving care in clinics.

To access the methods and software, please visit the generalizability_g_form_IPW and ExtendingInferences GitHub repositories.

Patients with chronic health problems, such as diabetes, often need to change treatment plans over time to improve their health. To help with this process, doctors can monitor patients' health through follow-up clinic visits and lab tests. Doctors may also suggest changing a treatment plan in response to visits or lab test results. When a treatment plan changes in this way, it's called a dynamic treatment plan. In this study, the research team developed and tested new statistical methods to learn how dynamic treatment plans and choices about follow-up care affect patients' health. These methods use electronic health records, or EHRs. Using EHRs is helpful because they have data on

• What treatments patients have received over time
• How treatments have affected patients' health
• Follow-up information such as lab test results

But the data may differ for patients based on when and why they go to the doctor. These differences make it hard for researchers to accurately know the effect of dynamic treatment plans across many patients.

To access the methods and software, please visit the simcasual R Package.

Clinical prediction models, or CPMs, are statistical models that can predict a patient's risk for a specific event, such as a health problem, adverse effect, or even death. To create a CPM, researchers use a single data set, such as data from a clinical trial. To find out whether the CPM accurately predicts risks for patients who weren't part of the original data, researchers can test the CPM with other data sets. This testing can help researchers know if the CPM is accurate for patients from different backgrounds and whether it can be used to make health decisions. But few CPMs have been tested with other data sets.

In this study, the research team used other data sets to look at how well CPMs for heart disease predict patients' risks. They also looked at how to improve CPMs.

To access the software and methods, please visit the Tufts Race CPM Registry.

Clinical trials study the effects of medical treatments, like how safe they are and how well they work. But most clinical trials don't get all the data they need from patients. Patients may not answer all questions on a survey, or they may drop out of a study after it has started. The missing data can affect researchers' ability to detect the effects of treatments.

To address the problem of missing data, researchers can make different guesses based on why and how data are missing. Then they can look at results for each guess. If results based on different guesses are similar, researchers can have more confidence that the study results are accurate. In this study, the research team created new methods to do these tests and developed software that runs these tests.

To access the sensitivity analysis methods and software, please visit the MissingDataMatters website.

Sometimes a study can get better results using data from different sites. In these cases, researchers may want to share patient data, including personal and private information such as dates of birth and addresses. However, researchers may not want to share data across sites because of worries about patient privacy. Some statistical methods can change patients' sensitive individual data into summary data that hides individuals' personal information. These privacy-protecting methods, or PPMs, make it safe to share data across sites. But researchers don't know if PPMs produce accurate results.

In this study, the research team compared combinations of PPMs with methods that use patients' individual data.

To access the methods, software, and R package, please visit the distributed GitHub.

Comparative effectiveness research compares two or more treatments to see which one works better for certain patients. Researchers often use data from patients' electronic health records to compare different treatments. This study addresses some problems that can arise from this practice. In some long-term research studies, researchers use data collected when patients in the studies see their doctors. Regularly scheduled doctor visits, called well visits, include yearly checkups or periodic blood pressure checks. Other doctor visits, called sick visits, occur when a patient feels sick or needs special care.

Well and sick visits can produce different types of health record data. In addition, test results at sick visits may be different from results at well visits. Using data from sick visits may inappropriately influence, or bias, a study's results. Also, patients may go to the doctor more often when they have symptoms or chronic health problems. Researchers may then collect more data from these patients than they collect from the healthier patients. Unequal amounts of data per patient make it harder to compare treatment results.

For this study, the research team created three tests to find if data from sick visits lead to bias in a study's findings. The team also compared standard and newer statistical methods for analyzing data that include sick visits. Researchers designed the newer methods to reduce bias from data obtained at sick visits. With less biased results, doctors can be more certain about which treatment worked better for certain patients.

Researchers can use data from health registries or electronic health records to compare two or more treatments. Registries store data about patients with a specific health problem. These data include how well those patients respond to treatments and information about patient traits, such as age, weight, or blood pressure. But sometimes data about patient traits are missing.

Missing data about patient traits can lead to incorrect study results, especially when traits change over time. For example, weight can change over time, and the patient may not report their weight at some points along the way. Researchers use statistical methods to fill in these missing data.

In this study, the research team compared a new statistical method to fill in missing data with traditional methods. Traditional methods remove patients with missing data or fill in each missing number with a single estimate. The new method creates multiple possible estimates to fill in each missing number.

To access the methods, software, and R package, please visit the SimulateCER GitHub and SimTimeVar CRAN website.

Electronic health records store a lot of data about a patient. These data often include age, health problems, current medicines, and lab results. Looking at these data may help doctors treating patients after a trauma predict how likely it is that they will respond well to a treatment and survive. This information can help doctors make better treatment decisions. But first, researchers need to figure out how to combine and analyze data to make accurate predictions. In this study, the research team created new statistical methods to combine data from patient records. They used these methods to predict patient health outcomes. Then the team used health record data collected from patients in hospital trauma centers to test their predictions.

To access the methods and software, please visit the following GitHubs:

• origami
• varimpact
• opttx

In randomized controlled trials, or RCTs, researchers assign patients by chance to different treatments to compare the benefits and harms. In RCTs, researchers have a high level of control over how patients receive treatment. RCTs often take place in research clinics with staff who monitor how patients follow treatment plans.

Pragmatic RCTs, or pRCTs, take place where patients typically receive treatment, such as a regular clinic. pRCTs can help capture the real-world effects of treatment but determining whether a treatment works can be hard in pRCTs. Also, no clear guidance exists about how to collect and analyze data from pRCTs. Some kinds of analysis are better for helping researchers focus on what's important to patients.

In this study, the research team created guidance for collecting and analyzing data in pRCTs so that results reflect what matters to patients and researchers.

To access the methods and software, please visit the following Github repositories:

• CDP-analysis-2018
• GFORMULA-RCT-SAS
• IV-Bounds
• CHARM_reanalysis
• Adherence_LRCCPPT