Monthly Abortion Provision Study
The Monthly Abortion Provision Study produces national and state estimates of the number of abortions provided within the formal health care system in the United States. This ongoing project reveals current trends and aims to put timely data in the hands of policymakers, advocates and providers.
What we're tracking
The Monthly Abortion Provision Study collects data on procedural and medication abortions provided at brick-and-mortar health facilities (such as clinics or doctor’s offices), as well as medication abortions provided via telehealth and virtual providers.
The data currently show:
The number of abortions provided by the US health care system (first tab)
Change in the number of abortions since 2020 (second tab)
In future months, this study will also provide estimates of the proportion of abortions that are obtained by patients traveling across state lines for care and estimates of abortion incidence by gestational duration.
For more than 50 years, Guttmacher has collected what is widely considered the most complete data on US abortion incidence and patient characteristics through our Abortion Provider Census and Abortion Patient Survey. The Monthly Abortion Provision Study complements Guttmacher’s in-depth efforts with a methodology adapted to the post-Roe policy environment, which has been characterized by rapid increases in state abortion restrictions, overburdened clinics, and heightened security risks for patients and providers. Monthly surveys collect data that—combined with statistical modeling—allow us to produce timely estimates of abortion caseloads and patients’ gestational duration and state of residence. These data provide policymakers, advocates, providers and the public with information on how shifts in policy influence abortion care, and they reveal trends over time that help to anticipate the impact of future policy changes.
How to read the data
The first graphic above shows the estimated number of abortions provided in each state for each month since January 2023. Abortions are counted as having been provided in the state in which a patient either had a procedure or where pills were dispensed. The graphic shows the median estimate for each month along with a range that describes the uncertainty: A bigger range means that estimate is more uncertain for a specific state and month, while a smaller range indicates that the estimate is more precise.
The second graphic shows the change in the number of abortions provided in each state in the year to date, as compared with estimates from a comparable period in 2020, as well as the percentage change from 2020. Data from 2020 is from our most recent Abortion Provider Census, which gathers comprehensive data on US national and state-level abortion incidence and care; to estimate a comparable period, we assume that abortion caseloads for January–June 2020 were equivalent to 50% of the abortion caseload for each state in that year.
The uncertainty intervals associated with each estimate quantify our confidence that the true value falls within a particular range. The 90% uncertainty interval describes the range in which our model is 90% certain the true value lies; the 50% uncertainty intervals are expected to contain the true value about half of the time. Our best estimate lies in the center of the range, but it is important to take the full range into account when interpreting these numbers. The uncertainty intervals are wider when there is less certainty in our estimates, such as when fewer data are available for a particular state or month, or when trends are particularly unexpected or extreme. In each round of surveys, we collect additional retrospective data, so estimates become more precise and the uncertainty intervals narrow over time.
Measuring the number of abortions that occur in the United States has always required some amount of estimation—and thus uncertainty—because of a number of complicating factors (including abortion stigma and incomplete reporting), and these challenges have only increased post-Dobbs. The Monthly Abortion Provision Study seeks to mitigate these challenges by combining data collected directly from abortion providers with a statistical modeling approach that helps us estimate caseloads at facilities for which we do not have data and helps us to better quantify our uncertainty. More specifically, we use a Bayesian hierarchical model to combine data from sampled providers with many years of detailed facility-level historical data on abortion caseloads from the Abortion Provider Census.
This study does not collect data on self-managed abortions, which we define as abortions occurring without in-person or virtual contact with the formal health care system. Some pregnant people might obtain abortion pills from a community support network or websites that source pills from outside the United States, including some that provide clinically supported care. Other people may use alternative or dangerous methods. Abortions in any of these categories are not reflected in Monthly Abortion Provision Study estimates. However, medication abortions that are prescribed via telemedicine by a doctor licensed in the United States and mailed to a patient in a state that allows telemedicine provision are included.
We update prior estimates monthly using data from a new sample of facilities, all of which provide data for the current month, as well as for past months in the current calendar year. As more data are collected, our estimates for past months get more accurate, so estimates may shift and uncertainty intervals will shrink. If you want to access the estimates as of the time they were published in any given month, they are available on the OSF website.
Because the confidentiality of providers and patients is paramount, we never release facility-level data. However, we encourage the use of our state-level and monthly estimates for further analyses. Data sets with all our estimates are available on the OSF website, including data sets with samples of posterior distributions to help researchers quantify uncertainty when using these data. We also publish the underlying model that produces our estimates and welcome feedback on it. We are continually updating and improving this model and will credit any researchers who make substantive contributions.
How we collect the data
The Monthly Abortion Provision Study leverages an innovative methodology—monthly collection of data from samples of abortion providers—to produce timely estimates of abortion care provided in the formal health care system. The primary aim of the study’s design is to meet the need for up-to-date data on the impact of rapidly changing state abortion policies while minimizing the burden we place on providers. Estimates are generated by a statistical model that combines data collected from monthly samples of providers with historical data on the caseload of every provider in the United States; as more data is collected each month, estimates for past months become more precise.
Monthly Abortion Provision Study estimates are based on data from two groups of facilities known to provide abortion in the United States. Facilities that play a particularly important or unique role in provision (e.g., because they border a state with an abortion ban or because they provide a large share of abortions in the state) are surveyed every month; the rest of our data come from samples of all other known US abortion providers. A new random sample is drawn each month, and facilities are asked to provide data on caseloads for all prior months in the current calendar year. All sampled facilities are also asked to provide data on patients’ state of residence, gestational duration and a rotating policy-relevant topic.
The rotating question is designed to provide timely information relevant to potential or recently enacted policy changes or gather information about aspects of care for which there is little recent research. For example, in early 2023, when the Alliance for Hippocratic Medicine v. U.S. Food and Drug Administration case first threatened the availability of mifepristone in the United States, the Monthly Abortion Provision Study asked providers about their use of other drug regimens for medication abortion. Other rounds of the survey have collected data on the proportion of abortions provided using medication and has asked providers about challenges their facilities are facing that they feel need more attention.
The Guttmacher Institute’s Abortion Provider Census (APC), fielded every three years since 1974, is the most comprehensive data collection effort on abortion provision in the United States. It collects data on abortion incidence and care from every known abortion provider in the United States. Each census is a sustained effort requiring a long fielding period, significant follow-up and substantial time and effort from abortion providers themselves. For instance, the 2019–2020 APC was fielded for a total of 15 months, during which time the fielding team conducted 7,131 phone calls, emails and faxes, and the resulting estimates were not available until mid-2022.
The Monthly Abortion Provision Study was designed to complement the APC and other data collection efforts by producing a slimmer portfolio of data but at a much quicker pace. Its design is responsive to the needs of policymakers, advocates and other stakeholders in the fast-changing landscape of US abortion policy and provision and attempts to minimize the time required of staff at surveyed facilities.
#WeCount is a collaborative national abortion reporting effort led by the Society of Family Planning that documents the changes in abortion volume by state. #WeCount aims to survey all clinicians providing abortions in the United States and has achieved high levels of participation from providers in many states. It has played a critical role in documenting the immediate impacts of the Dobbs ruling and will continue to track shifts in abortion volume over time. In states where participation has been significantly lower, data from known missing providers are imputed.
The Monthly Abortion Provision Study is based on data from samples of providers, which we combine with extensive historical facility-level data on variations in caseloads over time. The method is intended to minimize the burden on facilities (many of which are facing major upticks in demands on their staff in the wake of the restrictions imposed as a result of the Dobbs decision) and to reduce potential for nonresponse bias. The Monthly Abortion Provision Study also differs from #WeCount in that we systematically collect additional data beyond abortion counts, including state of residence, gestational duration and a rotating topic relevant to the current policy landscape. We present uncertainty ranges alongside our median estimates to highlight the uncertainty inherent to any abortion data collection effort, and when available, we present our estimates side by side with #WeCount’s estimates.
Protecting the confidentiality and privacy of providers and patients is our highest priority. Guttmacher employs stringent security measures to protect abortion-related data and the individuals who provide them. We never release information that could identify a specific facility or individual, and our estimates reflect only aggregate data. We are grateful to providers for trusting us with data from their facilities, and we continue to work to uphold that trust.
We are profoundly grateful to the staff at abortion providing facilities who provided data to this study—for their participation in this work, as well as for the care they provide to patients every day. We are also grateful to Marielle Kirstein, who helped conceptualize and develop this project, and to our fielding team (Mariah Menanno, Lauren Mitchell, Aisiri Murulidhar and Cici Osias), who work tirelessly to collect and check the data on which this study relies.
The Monthly Abortion Provision Study is supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development, the Office of Research on Women’s Health, and the Office of Behavioral and Social Sciences Research of the National Institutes of Health, under award number R61HD112921. The content is the sole responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.