Experiences of women who travel for abortion: A mixed methods systematic review

Objective To systematically review the literature on women’s experiences traveling for abortion and assess how this concept has been explored and operationalized, with a focus on travel distance, cost, delays, and other barriers to receiving services. Background Increasing limitations on abortion providers and access to care have increased the necessity of travel for abortion services around the world. No systematic examination of women’s experiences traveling for abortion has been conducted; this mixed-methods review provides a summary of the qualitative and quantitative literature on this topic. Methods A systematic search was conducted using PubMed, Embase, Web of Science, Popline, and Google Scholar in July 2016 and updated in March 2017 (PROSPERO registration # CRD42016046007). We included original research studies that described women’s experiences traveling for abortion. Two reviewers independently performed article screening, data extraction and determination of final inclusion for analysis. Critical appraisal was conducted using CASP, STROBE, and MMAT checklists. Results We included 59 publications: 46 quantitative studies, 12 qualitative studies, and 1 mixed-methods study. Most studies were published in the last five years, relied on data from the US, and discussed travel as a secondary outcome of interest. In quantitative studies, travel was primarily conceptualized and measured as road or straight-line distance to abortion provider, though some studies also incorporated measures of burdens related to travel, such as financial cost, childcare needs, and unwanted disclosure of their abortion status to others. Qualitative studies explored regional disparities in access to abortion care, with a focus on the burdens related to travel, the impact of travel on abortion method choice, and women’s reasons for travel. Studies generally were of high quality, though many studies lacked information on participant recruitment or consideration of potential biases. Conclusions Standardized measurements of travel, including burdens associated with travel and more nuanced considerations of travel costs, should be implemented in order to facilitate comparison across studies. More research is needed to explore and accurately capture different dimensions of the burden of travel for abortion services on women’s lives.


Introduction
The World Health Organization estimates that one in four pregnancies ends in abortion [1], yet access to this essential service is limited globally. Abortion is legally permitted only in cases where a woman's life is in danger or prohibited altogether in 66 countries; however, even in the 56 countries where abortion is technically available on request, restrictive abortion laws can make accessing abortion difficult [2][3][4]. Conscientious objection, the refusal to participate in abortion services because of religious, moral, philosophical or ethical beliefs, also affects access to abortion. Conscientious objection is legal in 21 European countries [5], including Italy, where 70 percent of gynecologists are conscientious objectors [3], and is practiced in the United States where it is federally protected under the "Church Amendments" [6]. In the United States, 90 percent of counties do not have access to an abortion provider [7]; such limited accessibility often necessitates travel for abortion services.
Traveling for abortion is not a new phenomenon; Irish and Canadian women have been traveling to the United Kingdom and the United States, respectively, to access abortion services since the 1960s [8]. More recently, restrictive measures including Targeted Regulation of Abortion Providers, or TRAP laws, have limited abortion access in many parts of the United States; one half of states have experienced a decline in the number of abortion facilities in the past 5 years, with some regions experiencing decreases up to 18 percent [7]. A recent study identified 27 major US cities that lacked a publicly advertised abortion facility within 100 miles, indicating that women across the United States must travel long distances for abortion services [9]. Even within states, there are large variations in abortion provider access, indicating spatial disparities that reflect a lack of access for women in rural areas [10]. With such limited provider options for women close to home, travel is often a necessary step in obtaining abortion services, but it does not come without costs. Studies from several states have reported negative effects related to abortion travel, including increases in travel time, transportation and childcare costs, stigma resulting from the need to disclose the abortion to others, and delays in care [11][12][13][14][15].
Given increasing limitations on abortion access in the United States, Europe and elsewhere [2,16], studying women's experiences of traveling for abortion is more important than ever. Despite an increase in research on this topic in recent years, there has been no comprehensive analysis of travel for abortion services. This paper aims to systematically examine the breadth and depth of the published literature on women's experiences traveling for abortion services by assessing different methodological approaches used and highlighting the importance of looking at travel in this field. While we cannot establish an overall description of travel and abortion due to the inconsistencies in methodologies and outcomes of the reviewed studies, we aim to assess how this concept has been explored and operationalized in order to show the need for consistent measurement of travel outcomes when accessing abortion and to provide recommendations for researchers on how to assess travel for abortion in future studies.

Search strategy
A PROSPERO protocol was registered for the review (#CRD42016046007) [17] and PRISMA guidelines were followed [18]. A search strategy was created by the first author, a clinical librarian (JBW), using keywords and controlled vocabulary, including MeSH, for the concepts of abortion and travel. The systematic search for articles on travel for abortion services was conducted on July 18, 2016 in PubMed, Embase, Web of Science, Popline, and Google Scholar. As this is a rapidly growing area of research, we re-ran our search on March 20, 2017 to ensure that new publications were included in the review. No date or language limits were used and unpublished and grey literature were not included. Detailed search strategies can be found in Appendix 1.

Study selection
Studies were excluded if they were not in English, did not contain original analysis, did not present data around travel for abortion, or involved a participant group other than women seeking or obtaining abortion services. Articles were double screened by two reviewers (JBW & RJ) based on title and abstract to determine if they met the inclusion criteria for full-text review. Articles without abstracts that appeared potentially relevant based on title were moved to final screening for further consideration. Full text screening was completed by two reviewers (JBW & RJ); each author screened all articles, and discrepancies were resolved by a third reviewer (CG).

Data extraction
Standardized forms were created to extract relevant data from the studies, including study setting, population, methodology, exposure and outcome measures, and results and conclusions related to abortion travel. Quantitative studies were categorized as retrospective if they relied on existing data sources such as hospital, clinic, or medical billing records; studies were considered to use prospective data collection if they enrolled abortion clients when they obtained or were seeking an abortion. We categorized studies as longitudinal if data was collected from participants at two or more time points; cross-sectional studies relied on only one data point per individual; ecological studies presented aggregate counts of the number of abortions or number of abortion clients at the facility, county, state, or national level. Studies were critically appraised using the STROBE checklist for quantitative studies, CASP checklist for qualitative studies, and MMAT checklist for mixed methods studies [19][20][21]. Data extraction was completed by research assistants and quality checked by two reviewers (JBW & RJ). Critical appraisal was completed independently by two reviewers (JBW & RJ) after completing consensus checks to ensure inter-rater reliability. Data extracted from quantitative and qualitative studies can be found in Appendices 2 and 3, and randomized critical appraisal data can be found in Appendices 4 and 5.
A thematic synthesis approach was used to analyze and synthesize the qualitative data. As our main research question was to understand how travel related to abortion was conceptualized and studied in the literature, we used an open coding process to develop salient themes. The results section from 12 qualitative studies and the qualitative results section from 1 mixedmethods study were uploaded into Dedoose, a web-based qualitative analysis software. Two authors (RJ & AR) conducted initial open coding on ten of the selected articles until saturation was reached, focusing only on results related to travel for abortion services. After substantial discussion, two authors (RJ & AR) grouped codes into second-order and first-order codes, with first-order codes representing overarching themes comprising both second-order and third-order codes (Appendix 6). These codes were applied to all qualitative findings. Any discrepancies between coding application were flagged and resolved through discussion. As all studies included in this review explicitly refer to study participants as female/woman, we use the words "female/woman" and the pronouns "she/her" throughout this paper. However, we acknowledge that some individuals who do not identify, as women are capable of pregnancy and may need timely access to safe abortion; while the perspectives of these individuals are not represented in the studies included in this review, findings may be relevant for these individuals as well.

Study selection
The systematic literature search yielded 664 articles. After excluding duplicates, 432 articles were screened for inclusion based on title and abstract. The full text of 120 articles was assessed for eligibility, and 73 were eliminated based on established inclusion and exclusion criteria. The search update on March 20, 2017 yielded an additional 12 studies for review. Fifty-nine studies were included in the final analysis: 46 quantitative studies, 12 qualitative studies, and 1 mixed-methods study (Fig 1). Table 1 describes the characteristics of the studies included in the review. The majority of studies (34 out of 59) used a cross-sectional design. Studies were published between 1975 and 2017, with the majority of studies (56%) published in the last five years, and a substantial number (18%) published more than 20 years ago. Over 90% of qualitative studies were published in the last two years. Studies represented findings from 11 countries, primarily from the United States (40 studies), Australia (five studies), the United Kingdom (five studies), and Canada (three studies). Of the quantitative and mixed-methods studies, 19 studies relied on secondary data collected from service records derived from patient charts, clinic billing data, or service records of callers to safe abortion hotlines [22] or national abortion funds [28]; 14 studies relied on secondary data sources such as state or national data on abortions. Eight studies relied on data collected from the Alan Guttmacher Institute (AGI) Abortion Provider Survey of health institutions and private physicians providing abortion services [29-30, 37-41, 55]. Twenty of the quantitative or mixed-method studies relied on primary data collected via selfadministered or interview-administered questionnaires with abortion clients. All of the qualitative studies relied on in-depth interviews with women seeking abortion or women who had obtained an abortion. Sample sizes ranged from 58 to 8,338 for quantitative studies that conducted primary data collection and 13 to 45 for qualitative studies.

Study characteristics
We created a "travel focus" variable for each study in order to judge its relevance to this review, indicated in Table 1. Studies labeled "primary focus" were studies that addressed factors related to travel as a critical component to analysis. For these studies, travel was considered either a main exposure, outcome (for analytic studies), or focus (for descriptive studies). For qualitative studies with a "primary" travel focus, the instrument guide was explicitly developed to gather information about women's experiences related to travel for abortion care. This category includes studies where the entire study population was women who traveled for abortions. Studies labeled "secondary focus" reported on results related to travel, but travel was not the primary outcome. For qualitative studies with a "secondary focus," travel may have been mentioned in descriptions of themes, but was not an overarching theme. Studies labeled "tertiary focus" only briefly mentioned travel: no specific results were presented and travel may have only come up in the discussion. Overall, 25% of studies (15) contained travel as a primary focus, 46% (27) had a secondary focus, and 29% (17) had a tertiary focus. These patterns were broadly consistent across quantitative and qualitative studies, study location, and date of publication, although the majority of studies with travel as a primary focus (67%) were published in the last five years. Studies with travel as a primary focus were concentrated in the United States, Australia, Canada, and England; all other settings discussed travel as a secondary or tertiary focus. Among the 40 US-based studies, 20% had travel as a primary focus, 53% had travel as a secondary focus, and 27% had travel as a tertiary focus.

Quantitative findings
Travel was measured and conceptualized in quantitative studies primarily as distance traveled and burdens related to travel, including travel costs incurred. As shown in Table 2, distance that women traveled for abortion services was measured in several ways within quantitative Experiences of women who travel for abortion: A mixed methods systematic review studies: 1) distance traveled in miles or kilometers was self-reported by patients or providers, 2) distance traveled in miles or kilometers was calculated using road networks, straight-line measurements, or geodesic formulas, or 3) time traveled was self-reported by patients. Similarly, burdens related to travel were conceptualized in quantitative studies in two principal ways. Eleven studies reported financial costs of travel, such as the cost of accommodations, gas, plane tickets, or other transportation costs. Five studies reported other burdens, including the need to arrange childcare, time away from work, and the need to disclose the abortion to others.
Quantitative studies (including one mixed-methods study) in this review considered a domain of travel as an outcome of interest (37 studies) or as an exposure (11 studies); one study explored travel as both an outcome and exposure [47]. Table 3 displays the range of approaches that studies took to understanding travel as a component of abortion seeking, often as a main outcome of a study that considered how far women needed to travel for abortion services [28,43,45,65] and sometimes as a secondary exposure, as in studies around home administration of medical abortion [51], abortion provider preference [59], or provider availability and abortion demand [23]. Many studies examined travel by measuring the amount of women who traveled out of state for abortion services [28-30, 39-40, 53]. Women who presented at Scotland facilities beyond the local gestational limit but whose pregnancies ended in abortion (assumption is that they needed to travel to procure the abortion) Of the 267 women, 18.7% (50) proceeded to abortion by traveling to England. Women who presented beyond 20 weeks had 6.37 times higher odds of continuing the pregnancy than those who presented in the 16th week (all women above 20 weeks would have had to travel for abortion services).
(Continued ) Time period (1983-1984 vs. 1993-1994) Travel distance to abortion clinic Change in travel distance to abortion clinic from 1983-1984 and1993-1994 Using state level data, calculated one way distance in miles between the abortion patient's resident and the location of the provider Rural women traveled farther for an abortion in 1993-1994 than in 1983-1984, and this difference was greater among older women.
Ellerston 1997 [27] Before and after the implementation of parental involvement laws (Continued ) Travel time to clinic (dichotomized as less than or equal to 3 hours or more than 3 hours), open ended responses to delays in reaching care related to travel categorized under "difficulty getting to the abortion facility" and "raising money for procedure and related costs" Women who obtained abortions after 20 weeks were more likely to have traveled more than 3 hours to reach a clinic (21% vs. 5%). Women who obtained abortions after 20 weeks were more than twice as likely than first trimester patients to report that difficult getting to the abortion facility slowed them down (27% vs. 12%) and spent more on transportation to the abortion facility. Approximately 10,000 women in the 6 months before the debate and passing of HB2 lived >200 miles from a Texas clinic providing abortions; this increased to 290,000 after HB2 restrictions began to be enforced. More than twice that many woman will live >200 miles from a Texas clinic when the ambulatory surgical center requirement goes into effect.
(Continued )    Studies that relied on self-administered or interview-administered surveys with abortion clients tended to focus explicitly on women's experiences related to travel for abortion care and conceptualized travel as the distance to the clinic along with additional travel-related burdens such as transportation and accommodation costs, time, and impacts on delays to care. These studies were all convenience samples that recruited participants from abortion clinics or provider settings. Studies that relied on state or national data were primarily focused on modeling the distance between counties of residence and the nearest abortion provider or the number of women who travel across state or national lines for abortion services. These studies relied on national or state surveillance reports from the study time period, or data on the number of abortions reported by the AGI Abortion Provider Survey, a comprehensive survey of all known abortion providers in the United States. Studies that relied on service records were mainly focused on the straight-line distance that clients traveled for their service and the number of out of state clients. These studies used all eligible records from a specified time frame (usually a particular year or range of years). For near-limits and turnaways, the most common delay was travel and procedure costs; these costs were higher for turnaways compared to those who obtained abortions in the first trimester. Near-limits traveled farther to reach a clinic than those who terminated in the first trimester and those who were turned away. Travel distance Whether women returned for abortion procedure, number of days between consultation and procedure visit Association between travel distance and returning for abortion procedure Distance traveled between woman's residential zip code and the facility where she attended the consultation visit, calculated using Stata's traveltime3 command 58% of women traveled less than 25 miles one way to the clinic, 13% traveled 25 to 49 miles, 21% traveled 50 to 100 miles, and 8% traveled more than 100 miles. Overall, 19% of women did not return to a clinic for an abortion procedure after their consultation. Distance traveled was not associated with return for an abortion visit. https://doi.org/10.1371/journal.pone.0209991.t003 Experiences of women who travel for abortion: A mixed methods systematic review Findings from studies included in this review suggest that the limited availability of abortion providers, insurance restrictions, as well as gestational age and other legal restrictions result in women needing to travel long distances for abortion services, often crossing state or country borders to seek care [27,28,31,33,45,50,60]. Studies that rely on aggregated countylevel data on the number of abortions found that abortion rates are lowest in counties that are farther from counties with abortion providers or have no abortion provider [16,23,42,46,58]. Studies that relied on state-level or national data described the phenomenon of women needing to travel out of state or to a different province for abortion services [28-30, 39, 40, 53] or to a different country [50,24]. Studies describe the substantial distances that women often need to travel in order to obtain abortion services; in these studies, many participants traveled over 50, 100, or even 200 miles to reach services [14, 15, 25, 26, 34-38, 41, 43, 47, 48, 52, 55, 57, 60, 62, 64]; rural women [25,26,43,45], women with gestational ages over 12 weeks [43,45,48], younger women [52], and women of lower socioeconomic status [47] were more likely to have to travel longer distances.
Studies describe a variety of additional burdens that traveling for abortion services places on women. Participants in these studies explicitly cite the cost of travel expenses [15,22,33,44,47,52,54,63,65], which include the cost of transportation, accommodation, childcare expenses, and lost wages as a barrier to reaching timely care when needing to travel for services. Some studies found that these burdens related to travel significantly negatively impacted participants' reproductive choices; needing to travel long distances for abortion services, particularly the logistical challenges and financial burdens associated with arranging travel, delayed women from accessing care in the first trimester [24,31,40,61], or prevented them from being able to obtain an abortion at all [24,61]. Two studies demonstrated the potential of telemedicine or at-home administration of medication abortion to reduce the burdens of Experiences of women who travel for abortion: A mixed methods systematic review traveling for services and increase access, particularly for women who live far away from an abortion provider [35,51].

Qualitative findings
All qualitative studies relied on in-depth interviews with women, either in person or via telephone/Skype, as their source of data. Of the twelve qualitative studies included in this review, two had travel as a primary focus [13,74], five considered travel as a secondary focus [11,12,[67][68]71], and five considered travel as a tertiary focus [65,[69][70][72][73]. All studies described women traveling either within or to countries where abortion is ostensibly legal albeit restricted (the United States, Scotland, Australia, and Canada); the context of many of these papers are to explore how laws or regional disparities in access affect women's experiences with abortion care, including their need to travel for services. Broad themes that emerged in how travel was conceptualized or discussed were the descriptive characteristics of travel, the burdens that needing to travel for abortion services imposed on women, the impact of travel on women's care and abortion method choices, and the reasons for travel. Key findings are presented below and themes from the qualitative studies, represented by commonlyseen terms, are presented in Fig 2. Codebooks can be found in Appendix 6.
Characteristics of travel. Almost all studies in this review contained descriptions of the modes of transportation women used when traveling for abortion services. Participants described traveling for abortion via airplane, private car, and public transportation. Needing assistance to arrange travel or obtain transport was described by participants in eight studies; for example, participants described needing to borrow money, borrow a car, or rely on a friend or family member for transportation to their appointments. Descriptions of characteristics of travel most commonly emerged in studies where travel was a secondary or tertiary focus.
Burdens of travel. Travel as a barrier to abortion care emerged as a theme in all studies, regardless of the travel focus of the study. Sub-themes related to the burdens of travel included logistical burdens, emotional burdens, cost, and time. Logistical burdens entailed the physical distance that women traveled to reach abortion providers, securing transportation, making childcare arrangements, and obtaining time off from work. Emotional burdens that were a direct result of travel included feeling uncomfortable and lonely while traveling alone for a procedure, feeling stressed from the need to figure out transportation and other logistics, as well as feeling stigmatized for the need to travel for routine medical care. Costs related to travel, including gas, hotel, childcare, and travel time, were mentioned in almost all studies.
In all but two studies, disclosure of having or needing an abortion, often to individuals that participants did not want to tell, were a direct result of the burden posed by needing to travel. Participants discussed how the need to secure time off of work, arrange childcare, or borrow money for travel or the procedure necessitated disclosing their decision to have an abortion to people at work and in their personal lives.
Impact of travel burdens on care. The impact of travel on women's abortion care was discussed in all but two studies. Most studies focused on the impact that travel had on women's choice of abortion method, obtaining care later than the woman would have wanted, and women's consideration of self-inducing. All but three studies discussed the impact of travel on restricting women's choice around their preferred method of abortion. For example, studies described how women chose surgical abortion over medication abortion because it would limit the distance they would need to travel to the clinic, the number of visits, as well as the possibility of experiencing abortion symptoms while traveling. In addition, studies also described how the burdens of travel necessitated staying overnight in order to facilitate their chosen method. In addition to limitations on method choice, the logistical aspects of travel described above often delayed women from getting to the clinic, resulting in restrictions to care based on gestational age. Four studies described how burdens imposed by travel were so great that they forced some participants to consider or attempt self-inducing to end their pregnancies.
Reasons for travel. Women's stated reasons for travel were often not explicitly addressed in the included studies. When stated, almost all reasons were framed in the contexts of increased legal restrictions that limited women's access to clinics or where residence in regions in which legal barriers to care necessitated travel, including presenting beyond gestational age limits for termination. Personal reasons outside of imposed restrictions rarely emerged; in one study, women reported traveling to the United States for abortion care because of perceived lack of safety of the procedure in Mexico (their country of residence) [65].

Critical appraisal
Studies were critically appraised using the STROBE checklist for quantitative studies [19], the CASP checklist for qualitative studies [20], and the MMAT checklist for mixed methods studies [21]. Each checklist assessed methodological quality of studies, including sampling, analysis, and bias considerations. Overall, study quality was good, and most studies adhered to principles of rigorous study design and analysis based on the checklist criteria. Older studies provided less information about recruitment, sampling, and data analysis techniques, resulting in better quality scores for studies published in the last 20 years.
For quantitative studies, quality issues arose around two main areas: participant flow and acknowledging bias within study design and results. Participant flow within studies was the largest area for improvement in reviewed studies. Many studies did not report numbers of participants at each stage of study (26 studies, 55%), explain how missing data was addressed (27 studies, 57%), or give reasons for non-participation at each stage (36 studies, 77%). Reporting on participant flow may help to ensure methodological quality as selection bias, missing data, and loss to follow-up can impact results, especially when discussing potentially sensitive topics around abortion experiences. 28% of studies did not acknowledge potential sources of bias in their studies, including possible bias within study design and the interpretation of results. Some studies may have worked to limit bias and did not report this; exhibiting transparency in these efforts would have resulted in higher critical appraisal quality scores. Older studies showed lower quality in these areas which may indicate a trend of acknowledging limitations with regard to study results; however, a majority of studies (68%) did not consider bias in study design, indicating room for improvement.
In qualitative studies, the principal quality issue involved a lack of consideration of the relationship between researchers and participants and the researcher's examination of their own role, biases and influence during formulation of the research question and data collection (57% of studies). Ethical issues and recruitment strategies were also areas for improvement, including details of how research was explained to participants, if ethics committee approval was sought, and if discussion around recruitment involving why participants were the most appropriate for these studies and why some participants did not choose to take part in the study took place.

Discussion
Given the different ways that travel has been measured and conceptualized in the literature, it is difficult to provide a cohesive summary statement of the conclusions of the studies in this review. However, taken together, these studies paint a picture of the reasons why women travel for services, barriers and delays they experience in traveling, the impact of this travel on their lives and reproductive choices, as well as possible solutions for reducing or eliminating this burden. Studies in this review suggest overall negative outcomes related to travel for abortion, including barriers related to monetary travel costs, time away from work, the need for childcare, and the need to disclose the abortion to others. Vulnerable populations were often more affected by travel, with younger women (including teenagers), women of color, and rural women traveling farther distances to access abortion services. Gestational age played a role as both an exposure and outcome related to travel in the reviewed studies: women at higher gestational ages often traveled farther distances to access abortion, and women whose limited access to abortion necessitated farther travel distances experienced delays that resulted in higher gestational ages or prevented them from obtaining an abortion altogether. Overall, travel impacted on women's access to abortion in multiple ways; causing delays, monetary costs, and other burdens.
Understanding and measuring the experience of traveling for abortion services is a complex challenge. A simple measurement of distance traveled will not suffice to capture the personal impact that distance has on the individual seeking an abortion. For example, if a woman does not have access to a personal car, then she is more likely to need to involve others in order to secure transportation, regardless of whether the distance is 15 miles or 50 miles. Similarly, there are relative measures of cost and additional burdens that women experience as a result of their travel for abortion services, some of which can be measured quantitatively and others qualitatively. While travel was approached in a range of ways and with a varying level of focus across the studies in our review, it was notable that there was no consistent measure for abortion related travel. This measurement inconsistency makes it difficult to compare travel outcomes across studies. Even a straightforward measurement like distance traveled was calculated by study authors using three different types of mathematical concepts. No pattern for use of travel measurement type (i.e. distance traveled by straight line, distance traveled by road network, time traveled) existed across studies: studies with travel as a primary, secondary, and tertiary focus showed equal amounts of each measurement.
Inconsistent measurement of travel is not unique to abortion related studies; the studies in this review reflect the wider healthcare field in terms of the variety of travel distance measurements used [75]. Standardizing measurements for abortion related travel, however, could have very real implications for policy planning and litigation, and researchers focusing on abortion should make efforts to address this issue. Standardized measures for travel related to abortion should be widely implemented to give an accurate, generalizable picture of the burden of travel for abortion on women's lives. The WHO recommends using travel time, rather than travel distance, as a measure of accessibility [76]; only six out of the 47 quantitative and mixed methods studies in this review used this measurement. Straight distance (Euclidean) and geodesic distance measurements may not accurately capture distance traveled, especially for women in rural areas who are often the target population in studies focusing on abortion. Recent studies, including 12 studies in this review, have started using road distance; with the increasingly worldwide coverage of Google Maps, researchers can use commands within Stata and similar software like Redivis to more easily estimate road distance and travel time measurements. Researchers may also look to recent real-world examples to create standardized measures of reasonable distances to travel for healthcare services; for example, the Veterans Access, Choice, and Accessibility Act of 2014, recently extended in 2017, allows veterans located farther than 40 miles (road distance) from the closest VA facility to seek care elsewhere [77].
Similarly, the variable of financial travel cost was not standardized in the reviewed studies. Financial costs were categorized in a variety of ways (e.g. $5-10, under $50, under $100). Existing national US data on out-of-pocket financial costs for abortion services [78] could be used to develop standardized cost measurements that would allow for comparison across studies in the United States and beyond. Financial costs for travel were rarely presented in the context of local, national, or individual expenses; national data on out-of-pocket health care expenses could be used as a standardized measurement of comparison to give context to women's financial burden of travel. In the United States, this data is measured in the Medical Expenditure Panel Survey as 1) the percentage of people with health expenses that had out of pocket expenses, 2) out of pocket expenses as a percentage of overall expenses, and 3) the average out of pocket expense per person [79]. Other burdens related to travel, such as time away from work and the inability to keep one's abortion confidential, were reported more often in qualitative studies. Qualitative research distinctly provides an opportunity to center the individual's experience of travel and to illustrate the complexity of experiences traveling for abortion services, and the interdependency of burdens in a way that quantitative instruments are not able to measure. More research should be done in this area to explore and accurately capture different dimensions of the relative burden of travel on women's lives. Additionally, qualitative results on other burdens of travel, such as forced disclosure and impact on decision-making, should be used to inform quantitative research, including data collection instruments. Most of the quantitative studies in this review relied on service records, or state/national data; as a result, analyses were often limited to calculating travel distance between clients' counties of residence and the location of the nearest clinic or provider. Future studies that directly survey women on their experiences related to travel may allow for a more nuanced and complete picture of the burdens related to traveling for abortion services to be documented in the published literature.
The majority of the studies in this review (40) are US-focused. There are several key differences between the United States and low-and middle-income countries (LMICs) in terms of travel distance and out-of-pocket cost for healthcare, including abortion care. Legal landscapes in Sub-Saharan Africa, the Middle East, Asia, and Europe often make travel for abortion necessary and unsafe abortion more commonly practiced [1]. Women in Sub-Saharan Africa often must travel long distances for all reproductive healthcare services, and out-of-pocket costs often prevent them from seeking care [80]. Despite some potential similarities in the need to travel for abortion care, it may be difficult to generalize the results of this review to LMIC settings where the legal and healthcare payment contexts are drastically different. Travel for abortion in developing countries, compounded with other barriers to care, could have a disproportionate impact on women's health compared to countries where abortion is legal; additional studies should explore this as a dimension of access. As access to abortion (and health care in general), should be within reach for all individuals, regardless of context, adaptations of recommended measures, such as travel time, could highlight disparities and barriers to abortion access that women all over the world face.

Limitations
There are some potential limitations of this systematic review. The majority of studies in the review are US-based which may make the findings difficult to generalize, particularly to settings where abortion is not legal. It may also be difficult to standardize and compare travel distance and costs globally; more research from non-US contexts is needed to address this issue. The studies in this review represent a range of designs, including modeling studies, those that contain national and state data, and studies involving discrete groups of women who travel for abortion. For this reason, comparisons across studies may be difficult, but because of the lack of available literature, all types of studies around travel for abortion were included to construct a complete summary of the research on this topic. Finally, the populations in the reviewed studies were women who traveled for abortions; there are no studies about women who did not make it to clinics or who chose to self-induce. Recent research shows that this may make up a considerable portion of women seeking abortions [81][82]; the experiences of these women around travel for abortion would contribute substantially to our understanding of barriers to abortion care.

Conclusion
This systematic review synthesizes the literature on the experiences of women who travel for abortion, including 46 quantitative studies, 12 qualitative studies, and one mixed methods study. Travel was categorized as a primary focus for 15 studies, a secondary focus for 27 studies, and a tertiary focus for 17 studies. Quality of studies was generally high, with participant flow and potential biases identified as areas for improvement. Travel distance, cost, and time were identified as burdens of travel for abortion care. Future studies should consider the use of standardized travel distance and cost measurements and continue to explore additional dimensions of the burdens of abortion related travel on the lives of women who need abortions.