R E S E A R C H A N D P R A C T IC E

Denial of Abortion Because of Provider Gestational Age Limits in the United States | U s h m a D. U p a d h y a y , P h D , T ra c y A. W e itz , P h D , R a c h e l K. J o n e s , P h D , R a n a E. B a ra r, M P H , a n d D ia n a G re e n e F o s te r, PhD

Objectives. W e e x a m in e d th e factors in flu e n c in g d e la y in seeking a b o rtio n and th e o u tc o m e s fo r w o m e n d e n ie d a b o rtio n care because o f g e s ta tio n a l a g e lim its a t a b o rtio n fa c ilities.

Methods. W e c o m p a re d w o m e n w h o p re s e n te d fo r a b o rtio n care w h o w e re u n d e r th e fa c ilitie s ‘ g e s ta tio n a l a g e lim its and received an a b o rtio n (n = 45 2 ) w ith th o s e w h o w e re ju s t o v e r th e g e s ta tio n a l age lim its and w e re d e n ie d an a b o rtio n (n = 231) at 30 U S fa c ilitie s . W e d e s crib ed reasons fo r d e la y in seeking services. W e e x a m in e d th e d e te rm in a n ts o f o b ta in in g an a b o rtio n e ls e w h e re a fte r being d e n ie d o n e because o f fa c ility g e s ta tio n a l age lim its. W e th e n e s tim a e n a tio n a l in c id e n c e o f being d e n ie d an a b o rtio n because o f fa c ility g e s ta tio n a l age lim its.

Results. A d o le s c e n ts and w o m e n w h o did not reco g nize th e ir p re g n a n c ie s e a rly w e r e m o s t likely to d e la y seeking care. T h e m o s t c o m m o n reason fo r d e la y w a s h a vin g to raise m o n e y fo r tra v e l and p ro c e d u re costs. W e e s tim a te d th a t each y e a r m o re th a n 4 0 0 0 U S w o m e n are d e n ie d an a b o rtio n because o f fa c ility g e s ta tio n a l lim its and m u s t c a rry u n w a n te d p re g n a n c ie s to te rm .

Conclusions. M a n y state law s restrict a b o rtio n s based on g e s ta tio n a l a g e, and n e w law s are lo w e rin g lim its fu rth e r. T h e in ciden ce o f being d e n ie d a b o rtio n w ill likely increase, d is p ro p o rtio n a te ly a ffe c tin g y o u n g and p o o r w o m e n . [Am J Public Health. 2 0 1 4 ;1 0 4 :1 6 8 7 -1 6 9 4 . d o i:1 0 .2 1 0 5 /A J P H .2 0 1 3 .3 0 1 3 7 8 )

The majority of abortions in the United States are in the first trimester of pregnancy, but 8.5% (approximately 100 000) occur after 13 weeks gestation.1 Most women having second trimester abortions would have liked to have had the procedure earlier,2 and women report a number of delaying factors, including cost and access barriers and late detection of preg nancy.2-4 These delays can result in women being denied care because they present with pregnancies beyond an abortion providers gestational age limit and are unable to obtain an abortion elsewhere. (An abortion provider is a facility where abortions are performed.5) Little is known about how frequently this occurs and what happens to women denied abortion care.

The 1973 Supreme Court Roe v. Wade6 decision established the point of potential fetal viability as the threshold after which states could restrict womens access to abortion care as long as they allowed for exceptions to preserve the life and health of the pregnant woman. However, Roe v. Wade did not specify a gestational age for viability. Many states have established an upper gestational limit, most commonly after 24 weeks from a wom ans last menstrual period, and some states have done so without the required excep tions.7 At least 8 states have recently reduced or plan to reduce the upper gestational limit to 20 weeks, and 1 state to 18 weeks.8 Individual abortion providers can set their limits at lower gestational ages, and do so based on the availability of trained physicians, clinician and staff comfort, and facility regulations. According to a national survey of abortion providers, 23% offer abortions after 20 weeks gestation, and 11 % do so at 24 weeks.5 Because fewer providers offer abortion care after the first trimester, women must travel longer distances to obtain later abortions. Because later abortions are more complex procedures, often occurring over 2 or more days, they are also more costly; the average

charge for an abortion at 10 weeks is $543 compared with $1562 for an abortion at 20 weeks.5 Some women must also arrange for childcare, take time off work or other responsibilities, and incur transportation and hotel expenses; raising these funds results in additional delays.9

W e sought to describe the characteristics associated with being turned away because of provider gestational age limit, and the efforts such women make to obtain a desired abor tion. Additionally, we explored the factors associated with obtaining a desired abortion elsewhere. Finally, we estimated the incidence of women being denied an abortion in the United States because of provider gestational limits.

M ETHODS

We obtained the data for this study from 2 sources, the University of California, San Francisco (UCSF) Tumaway Study and the Guttmacher Institutes Abortion Provider

Census. Both studies were approved by their institutional review boards.

The Tumaway Study is a 5-year longitudi nal prospective study of women who receive an abortion and women who are denied an abortion because they present for care after the providers gestational limit. The study was designed to assess a variety of outcomes of receiving an abortion compared with carrying an unwanted pregnancy to term. From 2 008 to 2010, the Tumaway Study recruited women from 30 abortion providers across the United States. Only last stop providers were selected, defined as being more than 150 miles from a facility with a later gestational limit. They were located in 21 states distrib uted relatively evenly across the country. Women were recruited on a 1:2:1 ratio: women who presented up to 3 weeks over the providers gestational limit and were turned away (tumaways), women who presented up to 2 weeks under the limit and received abortions (near-limit abortion patients), and women who presented in the first trimester

S eptem ber 2 0 1 4 , Vol 1 0 4 , No. 9 | Am erican Journal o f Public Health Upadhyay et al. | Peer Reviewed | Research and Practice | 1 6 8 7

R E S E A R C H A N D P R A C T IC E

and received abortions (first trimester patients).

Women were eligible for participation if they sought an abortion within the gestational limits for each of the study groups, spoke English or Spanish, and were aged 15 years or older. Further details on recruitment and methods can be found elsewhere.10-12 After the baseline survey, study participants were con tacted for a follow-up phone interview every 6 months for 5 years. Tumaway Study data for this analysis were from the baseline (1 week after recruitment) and 6-month interviews.

To reduce losses to follow-up, researchers collected detailed contact information and participants preferred methods of communi cation and confidentiality protection prefer ences; they also called women after 2 months to confirm that the womans primary and secondary contact information was still valid. When participants could not be reached, re searchers called each day, for up to 5 days. If the participant still could not be reached, re searchers sent up to 3 follow-up letters by mail or e-mail (according to stated contact prefer ences) and continued to call at the same frequency for a maximum of 10 sequential days. To mitigate respondent burden and to compensate them for their time, participants received a $50 gift card to a large retail store upon completion of each interview.

Facility data from the Guttmacher Provider Census were used to estimate the incidence of being denied an abortion nationally be cause of facility gestational age limit. These data have been used to create national esti mates of a variety of abortion-related indica tors.5,13-16 In 2009, the Guttmacher Institute surveyed all US facilities known to have performed abortions. In May 2009, up to 3 rounds of questionnaires were mailed to all potential providers, and extensive phone follow-up was conducted. Of the 2 3 4 4 facili ties surveyed, 1024 responded to the mailed questionnaire, 501 responded during nonre sponse follow-up, and health department data were used for 451 facilities. A total of 1793 facilities reported providing at least 1 abortion in 2008. All facilities were asked the number of abortions performed and the maximum gesta tion at which abortion services were offered, and about other aspects of abortion care. Further details can be found elsewhere.5

M e a s u r e s

All measures were taken from the Tumaway Study. During the baseline Tumaway Study interview, participants were asked about soci odemographic characteristics, their reproduc tive histories, when they discovered they were pregnant, when they first considered an abor tion, and any difficulties they experienced accessing care. Access to abortion was oper ationalized by the distance women traveled to get to the study recruitment facility and the number of other facilities they called or visited before presenting at the recruitment facility. Distance traveled by road was estimated using the STATA module (StataCorp, College Station, Texas) TRAVELTIME17 which uses Google maps (Google, Mountain View, California) to geocode distance between participant and facility zip codes.

All participants were asked about the rea sons for delay in seeking an abortion. First, they were asked an open-ended question, Did anything slow you down and prevent you from getting to the [Recruiting site] earlier in your pregnancy? A few months after data collection began, additional closed-ended items about specific reasons for delay were added to the instrument, including: not knowing you were pregnant, not knowing where to go to get the abortion, figuring out how to get to the provider, travel costs, costs of the procedure, insurance coverage, and trouble deciding whether an abortion was wanted. Participants could select all that ap plied. Responses from the open-ended ques tion were coded by 2 of the authors and combined with the closed-ended item cate gories. W e conceptualized insurance delays as distinct from procedure cost delays, in that the former included administrative and logis tical problems, such as having to determine whether the procedure was covered among women with insurance or waiting for Medicaid- based coverage. Women who were not asked the closed-ended items and did not mention the issue in response to the open-ended question were coded as missing for that category. At baseline, and if they had not had an abortion at baseline, again at 6 months, tumaways were asked whether they considered obtaining an abortion elsewhere, whether they had obtained an abortion elsewhere, and what barriers they faced in accessing care.

D a t a A n a ly s is

We conducted the analysis in 3 parts. First, we compared the sociodemographic charac teristics and reproductive history and inten tions of first trimester patients, near-limit abortion patients, and tumaways. Because al most one third of participants, nearly exclu sively women living with their parents, did not know their household income, we also examined mothers education as a proxy for socioeconomic status. We used bivariate mixed-effects regression models that in cluded random effects for facility, presenting P values that adjusted for the clustering of participants within providers. We used mixed-effects logistic regression to assess group difference in binary variables, mixed-effects multinomial logistic regression for categorical variables, and mixed-effects ordinal logistic regression for ordered cate gorical variables. For continuously coded characteristics, we used mixed-effects linear models to assess differences in means among the study groups. W e also described the access-related barriers and compared these experiences by study group using mixed-effects regression models.

Second, we examined the factors associated with having an abortion after being denied one. We fit a multivariable mixed-effects logistic regression model to assess the characteristics associated with ultimately obtaining an abor tion among tumaways. The model included sociodemographic variables and reproductive history and intentions. Because of clustering of gestational age by site, both site gesta tional limit and the individuals deviation from the site gestational limit were entered into the model. Statistical significance was set at P < .0 5 for all comparisons and adjusted odds ratios (AORs), and 95% confidence intervals are reported. All statistical analyses were performed using STATA 12 (Stata Corp, 2011).

Finally, we estimated the incidence of being denied an abortion because of provider ges tational limit and projected the number of women affected nationally per year. We used data provided by 4 study providers (1 on the east coast, 2 in the midwest, and 1 on the west coast) that maintained records of all women denied care because of gestational limits to estimate the proportion of all clients

1 6 8 8 | Research and Practice | Peer Reviewed | Upadhyay et al. American Journal o f Public Health | Septem ber 2 0 1 4 , Vol 10 4 , No. 9

RESEARCH AND PRACTICE

who presented beyond the facilitys gesta tional limit. All 4 facilities performed at least 4 0 0 abortions, provided abortion care at the latest gestational age for at least 150 miles, and went to at least 13 weeks gestation, characteristics that were similar to abortion facilities in the larger sample of last-stop abortion facilities.

We applied this estimated proportion to the total number of women seeking abortions at last stop facilities based on the Guttmacher Provider Census data. Last stop facilities were those that performed > 4 0 0 abortions in 2008, provided abortion care at the latest gestational age for that state, but went to at least 13 weeks gestation, and were more than 150 miles from a facility in a bordering state that had a later gestational limit.

R E S U L T S

Among the 3045 women who were approached, 39.4% were interested in being interviewed semiannually for 5 years, and agreed to speak with UCSF researchers by phone. The most common reason for refusal was the time commitment required for par ticipation. Among these, 94.4% were eligible, gave informed consent to participate in the study, and were enrolled. After stratifying by study group, nonparticipants (among those consented) did not differ from participants on age or gestational age of the pregnancy at the time of enrollment. A total of 956 women completed the baseline interview: 273 in the first trimester group, 45 2 in the near-limit abortion group, and 231 in the tumaway group. Among the women enrolled, 92% were retained at 6 months, with no differential loss to follow-up among groups.

The sample was racially and ethnically di verse, with more than half (50.4%) being Latina or African American (Table 1). The majority were single and never married (78.8%), and most had previous children (59.2%).

Sociodemographic characteristics of first tri mester patients differed substantially from near-limit abortion patients. Near-limit abor tion patients were less likely to be aged 25 to 34 years, more likely to be multiracial or other race, less likely to have a college degree, less likely to be in the highest income category, and less likely to be employed. Near-limit abortion

patients discovered their pregnancies at later gestational ages than first trimester patients, and near-limit abortion patients were less likely to report difficulty deciding about the abortion.

There were few sociodemographic differ ences between tumaways and near-limit abor tion patients; tumaways were younger, less likely to be employed, and less likely to have children than were near-limit abortion patients. Most notably, tumaways discovered their pregnancies at later gestational ages than did near-limit abortion patients.

R e a s o n s fo r D e la y a n d A c c e s s B a r r ie r s

Among all causes of delay, tumaways were more likely than first trimester patients to report that each reason caused a delay except for difficulty deciding whether to have an abortion (Figure 1). Reasons for delay in cluded travel and procedure costs (36.5% among first trimester patients and 58.3% among tumaways), not recognizing the preg nancy (37.8% among first trimester patients and 48.1% among tumaways), insurance problems (20.3% among first trimester pa tients and 37.2% among tumaways), not knowing where to find abortion care (19.9% among first trimester patients and 33.5% among tumaways), and not knowing how to get to a provider (12.8% among first trimester patients and 29.8% among tumaways; all Rvalues < .05).

Between tumaways and near-limit abortion patients, there were no significant differences in reasons for delay. For women in both groups, the most common reason for delay was travel and procedure costs. Most responses to the open-ended questions did not specify which costs caused the delay: women com monly cited, money, and finances.

Near-limit abortion patients and tumaways reported a variety of additional life circum stances that did not fit the predeveloped cate gories of reasons, including (in no order) having to wait a while for an appointment, opposition from family or friends, being in jail, needing to obtain an ID or birth certificate, weather (ice storm, blizzard, or flooding), fear of protesters, difficulties getting time off work, and difficulties getting childcare. A few women cited problems with referrals; for example, 1 woman reported that she had to

wait a week before she could get an ap pointment at another provider, and by then she had also surpassed the new providers gestational limit.

Generally, near-limit abortion patients went to greater lengths than tumaways to obtain an abortion. Although because they were at later gestational ages, tumaways may have had fewer provider options (Table 2). Near limit abortion patients traveled greater dis tances than first trimester patients (30.5% vs 13.6% traveling > 100 miles, P < .0 0 1 ), and called (49.4% vs 34.9%, P < .001) and visited (51.9% vs 32.2%, P < .0 0 1 ) more providers. Near-limit abortion patients also traveled greater distances than tumaways (30.5% vs 19.5% P< .001) and were more likely to have visited other providers before presenting at the re cruitment site (51.9% vs 34.5%, P < .0 0 1 ).

F a c to r s A s s o c ia te d W it h O b ta in in g a n

A b o r tio n A f t e r B e in g D e n ie d O n e

Among the 231 tumaways, 48.5% said they did not consider having an abortion elsewhere after being denied one; however, among these women, over half (55%) said they still wished they could have had an abortion.

An additional 21.6% of tumaways said they considered having an abortion else where, but never obtained one. Among this group, the most commonly reported reason for not obtaining an abortion after being denied one were procedure and travel costs (85.4%), followed by not being able to find a provider who would do the abortion so late, not knowing where to go, or a belief that no services were available for their gesta tional age (54.8% combined), and not knowing how to get there (51.1%). One woman cited the burden of the combination of factors:

It was probably travel costs, procedure costs, not knowing who I would have to come with me on the four day adventure. I was at the point that there was no guarantee wherever I went.

Six months after recruitment into the study, 64 of the 231 tumaways (27.7%) had re ceived an abortion, and 5 women (2.2%) had had a miscarriage or stillbirth. Among all tumaways, 15 (6.5%) placed their children for adoption (9.3% among those who gave birth).

S eptem ber 2 0 1 4 , Vol 1 0 4 , No. 9 | American Journal o f Public Health Upadhyay et al. | Peer Reviewed | Research and Practice | 1 6 8 9

RESEARCH AND PRACTICE

TABLE 1-Sociodemographic Characteristics and Reproductive History of the Study Population by Study Group: United States, University of

California, San Francisco Turnaway Study, 2 0 0 8 -2 0 1 0

Characteristics Total % First Trimester

Patients (F) (n = 273) % P (F vs N) Near-Limit Abortion

Patients (N) (n = 452) % Turnaways

(T) (n = 231) % P (N vs T)

Gestational age (mean d) 163.1 77.2

Sociodemographics

<.0 1 189.7 212.7 < .01

Age, y

15-173 18.1 13.2 Ref 16.8 26.4 Ref

18-24 36.1 30.4 .94 39.4 36.4 .01

25-34 38.3 47.3 .03 36.3 31.6 .01

35-46 7.5 9.2 .19 7.5 5.6 .04

Race/ethnicity

Non-Hispanic White 36.9 42.5 Ref 34.7 34.6 Ref

Non-Hispanic Black 29.4 29.3 .1 29.4 29.4 .67

Hispanic/Latina 21.0 20.5 .15 20.4 22.9 .37

Multiracial/other 12.7 7.7 < .01 15.5 13.0 .76

Highest grade completed

< high school 19.6 15.8 .73 19.0 25.1 .09

High school diploma or GED 33.4 31.5 Ref 35.2 32.0 Ref

Some college, vocational training 39.3 41.4 .36 39.2 37.2 .82

College degree 7.7 11.4 .03 6.6 5.6 .84

Poverty status’1

< 100% FPT 33.6 29.7 Ref 35.8 33.8 Ref

100%-200% FPT 21.8 25.3 .1 21.5 18.2 .66

> 200% FPT 12.7 20.9 < .0 1 10.4 7.4 .37

Dont know household income 32.0 24.2 .62 32.3 40.7 .13

Employed

Unemployed 46.9 37.4 Ref 46.7 58.4 Ref

Part or full time 53.1 62.6 .01 53.3 41.6 < .01

Maternal education

< high school 15.1 20.5 .04 12.6 13.4 .9

High school 36.4 36.3 Ref 36.1 37.2 Ref

Some/grad tech, or college 18.0 13.9 .15 19.2 20.3 .91

> college grad 22.1 24.2 .67 21.9 19.9 .58

Dont know 8.5 5.1 .04 10.2 9.1 .63

Insurance status (n = 953)

None 28.8 28.9 .31 29.6 26.8 .86

Medicaid 43.8 39.2 .12 43.7 49.4 .2

Private/other 27.5 31.9 Ref 26.7 23.8 Ref

Marital status

Single, never married 78.8 75.1 Ref 79.2 82.3 Ref

Married 9.1 11 .18 8.0 9.1 .77

Separated, divorced, widowed 12.1 13.9 .61

Reproductive history and intentio ns

12.8 8.7 .11

Previous children (n – 954)

0 40.9 40.8 Ref 37.5 47.6 Ref

1 27.4 24.6 .13 30.4 24.7 .02

> 2 31.8 34.6 .93 32.2 27.7 .04

Continued

1690 | Research and Practice | Peer Reviewed | Upadhyay et a/. American Journal of Public Health | September 2014, Vol 104, No. 9

R E S E A R C H A N D P R A C T IC E

TABLE 1 – C o n t in u e d

P re v io u s a b o r s ( b e f o r e in d e x a b o r t io n ) (n = 9 5 5 )

0 5 4 . 3 5 3 . 7 R ef 5 3 . 3 5 7 . 1 R ef 1 2 7 . 3 2 5 . 4 .5 2 9 . 0 2 6 . 4 . 3 8 > 2 1 8 . 3 2 1 . 0 .3 9 1 7 . 7 1 6 . 5 .5 1

H o w d if fic u lt to m a k e t h e d e c is io n

V ery o r s o m e w h a t d iffic u lt 4 4 . 4 5 2 . 0 .0 1 4 1 . 2 4 1 . 6 .9 3 V ery o r s o m e w h a t e as y, n o t e a s y o r d iffic u lt 5 5 . 6 4 8 . 0 5 8 . 8 5 8 . 4

G e s t a t io n a l a g e w h e n d is c o v e re d p re g n a n c y (n – 9 5 4 ) , wk

< 1 0 6 5 . 9 9 9 . 3 R e f 5 7 . 6 4 2 . 6 R ef 1 1 – 2 0 2 4 . 9 0 . 7 < . 0 1 3 2 . 8 3 8 . 3 .0 1 > 2 0 9 .1 0 . 0 NA 9 . 5 1 9 .1 < . 0 1

N o t e . FPT – f e d e r a l p o v e rty t h r e s h o ld ; GEO – g e n e ra l e q u iv a le n c y d ip lo m a ; NA = n o t a p p lic a b le : P v a lu e c o u ld n o t b e c o m p u t e d b e c a u s e o f e m p ty c e ll. T h e s a m p le s iz e w a s n – 9 5 6 u n le s s in d ic a t e d . T h is a g e c a te g o ry in c lu d e s 1 w o m a n a g e d 1 4 y e a rs w h o w a s re c ru ite d e a rly in t h e s tu d y b e fo re t h e m in im u m e n r o llm e n t a g e w a s c h a n g e d t o 1 5 y e a rs . T h e FPTs w e re d e f in e d by t h e U S C en s u s ( 2 0 0 8 – 2 0 1 0 ) .

At the provider with the lowest estab lished gestational age limit in the study (10 weeks), 2 0 of the 21 women turned away (95.2%) eventually obtained an abor tion despite being more than 150 miles from another facility. W hen women from this 1 site were excluded, 21.5% of turnaways were able to obtain an abortion. Among turnaways who had an abortion, the majority (84.1%) found out about the providing facility from the original recruitm ent pro vider. An additional 7.9% reported learning about the providing facility from another

health care provider, whereas another 7.9% reported learning about it from other sour ces such as the Internet and the National Abortion Federation hotline.

Results of the multivariable model pre dicting the likelihood of obtaining an abor tion after being denied one demonstrated that women who were Latina (AOR = 0.12, 95% confidence interval [Cl] = 0.03, 0.56), who reported it was very or somewhat difficult to make the decision to have an abortion (AOR = 0.19, 95% Cl = 0.07, 0.49), and who were recruited at a facility with a later

gestational limit (AOR = 0.68, 95% Cl = 0.61, 0.77) were less likely to have an abortion after being denied one than were other women (Table 3).

Incidence of Being Denied Abortion Because of Gestational Limits

The providers in our study had gestational limits from 10 to 26 weeks, with a mean limit of 20 weeks. Based on data from 4 of the study facilities with complete records on women turned away, we estimated that facilities turned away an average of 2.0% of clients seeking care because they presented for care after the providers gestational limit.

Using the Guttmacher Provider Census, we estimated that there are 101 last stop pro viders across the United States. These pro viders have a total patient volume of about 263 917 per year. Applying the 2.0% turn- away rate, we estimated that in 2008 ap proximately 5278 women presented at pro viders but were denied an abortion in the United States because they were beyond the providers gestational limits. Based on the proportion of turnaways (at providers with limits at 13 weeks or more) who were able to obtain an abortion elsewhere (21.5%), we estimated that 4 143 women carried their unwanted pregnancies to term. These esti mates did not include women who were denied care for other reasons such as medical ineligibility, not having funds to pay for the abortion, or not having permission from a par ent (where parental consent was required).

Travel and Not recognizing D ifficulty Insurance Not know ing Not knowing procedure pregnancy* deciding problem s*** where to g e t h o w to ge t to a costs*** w hether to have care** provider***

abortion

Causes o f Delay

* P < . 0 5 ; * * P < . 0 1 ; * * * P < . 0 0 1 ; P v a lu e s r e fle c t d iffe re n c e s b e tw e e n firs t t r im e s t e r p a t ie n ts a n d tu rn a w a y s .

FIG U R E 1 – R e p o r t e d c a u s e s o f delay, by study group: U n ite d S ta te s , U niversity o f C a lifo rn ia ,

S a n Francisco Turnaw ay Study, 2 0 0 8 – 2 0 1 0 .

S eptem ber 2 0 1 4 , Vol 1 0 4 , No. 9 | American Journal o f Public Health Upadhyay et at. | Peer Reviewed | Research and Practice | 1 6 9 1

R E S E A R C H A N D P R A C T IC E

T A B L E 2 – A c c e s s t o A b o r t i o n b y S t u d y G r o u p : U n i t e d S t a t e s , U n i v e r s i t y o f C a l i f o r n i a , S a n

F r a n c i s c o T u r n a w a y S t u d y , 2 0 0 8 – 2 0 1 0

First Trimester Near-Limit Patients Abortion Patients Turnaways

Access Indicator Total % (F; n = 2 7 3 ), % P (F vs N) (N; n = 4 5 2 ), % (T; n = 2 3 1 ), % P (N vs T)

Distance to provider (n – 9 5 6 ), miles

< 5 0 5 9 .2 7 1 .4

5 0 – 1 0 0 1 7 .8 1 5 .0

> 1 0 0 2 3 .0 1 3 .6

No. of providers called (n = 9 4 7 ) No other providers 5 6 .2 6 5 .1

> 1 other provider 4 3 .8 3 4 .9

No. o f providers visited (n – 9 4 7 ) No other providers 5 8 .0 6 7 .8

> 1 other provider 4 2 .0 3 2 .2

Other women not included in our estimate were women who knew their gestational age, in quired about the facilitys gestational limit by phone, and never presented for care.

D I S C U S S I O N

Findings from this study suggest that in 2 0 0 8 m ore than 4 0 0 0 women carried un w anted pregnancies to term after they were denied an abortion because of provider gesta tional age limits. This study was initiated before the recent state abortion bans at 2 0 weeks gestation. Almost 15% of US w omen live in the states with such new legislation; thus, many m ore will be denied abortions in the coming years. These bans present an undue burden because, as dem onstrated in this study, many w omen do not realize they are pregnant until later in pregnancy and cannot travel to other states for abortion care. Children b o m from unintended pregnancies have multiple health consequences18′ 21 compared with children b o m from intended pregnancies. Additionally, women who raise children b o m from unin tended pregnancies have higher rates of eco nomic22 and educational2 1 disadvantages.

In this study, one of the primary reasons for delay in seeking an abortion was time spent raising the funds to pay for the procedure and travel. Once a w om an is beyond the first trim ester, raising the funds to pay for the abortion can lead to further delays and create a cycle of increasing cost and delay. Currently, in

< .0 0 1 5 1 .1 6 0 .6 < .0 0 1

1 8 .4 19.9

3 0 .5 1 9 .5

5 0 .6 5 6 .5

< . 0 1 4 9 .4 4 3 .5 .12

4 8 .1 6 5 .5

< . 0 1 5 1 .9 3 4 .5 < . 0 1

33 states and the District of Columbia, poor women have no access to Medicaid-funded abortions, except in cases of life endangerment, rape, or incest.24 Public financing and insurance coverage for abortion would have made pro cedures possible for many of the turnaways, and ability to pay while in the first trimester could have prevented some women from needing later abortions. These findings were consistent with those of Henshaw et al.25 who estimated that one fourth of women who would have had Medicaid-funded abortions instead gave birth when this funding was unavailable.

W e found that first-trimester patients were m ore likely to re p o rt difficulty deciding w hether to have an abortion than both near- term abortion patients and turnaways. This might be because first-trimester patients had few er o th er delays. It was easier for them to find a provider, raise the m oney, etc., so their definition of delay was relatively lower.

Our findings demonstrate a need to strengthen existing financial support and referral systems to ensure that women can be served elsewhere if they cannot be treated where they originally present for care. Referrals could be made immediately at the facility that denied care or via a phone consultation service. A few organizations, such as the National Abortion Federation and the National Network of Abor tion Funds, provide women with information about abortion providers nearest to them, in cluding the latest gestation at which abortions are available, as well as financial assistance.26,27

However, the financial support for these services is limited and privately donated, and many women are unaware of these resources or unable to access them.

Expanding the num ber of abortion facilities in underserved areas and enabling providers to raise their gestational limits would likely reduce out-of-pocket costs associated with travel, time off work, and childcare. Several factors influence how providers set their ges tational limits. An informal survey conducted by one of the authors in 2 0 0 7 among 74 second trim ester abortion providers found that the most commonly reported factors in de termining gestational limit were surgical skills and comfort (71%), state regulations (42%), and personal beliefs (37%).28 Potential strate gies for raising limits include investing in training clinicians to perform later procedures, ensuring these providers have ample institu tional and emotional support,29,30 and addressing the social stigma that they face.31,32

S tu d y L im ita tio n s

This study had a few limitations. First, the Turnaway study was limited to fewer than 1 000 women, and many w omen who were invited to participate declined. T he percentage of w omen who did not want to participate varied widely by facility, with the 10 top- ranking facilities achieving 6 0 % to 8 0 % agreem ent and the bottom 5 facilities having less than 3 0 % agreement. T he low recruitment rate among some facilities was likely because of the long-term demands of study participa tion. To assess the extent of this limitation, we compared the outcomes of women who w ent to high recruiting facilities with those who went to low recruiting facilities and found very little difference.10 W e also compared the dem o graphics of the w omen in our study to the demographics of women receiving abortions nationally, and found that they were very similar, with the exception that our participants were m ore likely to be in the second trimester. To the extent that those who refused to participate experienced different barriers to accessing care, this could have affected our findings. Second, we had high rates (16% 20% ) of missing data on reasons for delay because early in data collection w omen were not asked the specific questions. However, data were not differentially missing betw een

1 6 9 2 | Research and Practice | Peer Reviewed | Upadhyay et al. American Journal o f Public Health | Septem ber 2 0 1 4 , Vol 1 0 4 , No. 9

R E S E A R C H A N D P R A C T IC E

TABLE 3 – U n a d ju s t e d an d A d ju s ted Odds o f O b ta in in g an A b o rtio n A fte r B e in g Turned Away:

U n ite d S ta te s , U n iversity o f C a lifo rn ia , S a n Francisco Turnaw ay Study, 2 0 0 8 – 2 0 1 0

Odds of Having an Abortion After Being Turned Away

Unadjusted OR (95% Cl) Adjusted OR (95% Cl)

S o c i o d e m o g r a p h i c c h a r a c t e r i s t i c s

Age, y 1.02 (0.95, 1.08) 1.00 (0.91, 1.09) Race/ethnicity

Non-Hispanic White (Ref) 1.00 1.00 Non-Hispanic Black 0.43 (0.16, 1.15) 0.42 (0.14, 1.31) Hispanic/Latina 0 .1 8 ** (0.05, 0.62) 0 .1 2 ** (0.03, 0.56) Multiracial/other 0.36 (0.11, 1.21) 0.40 (0.10, 1.59)

Highest grade completed

< high school 0.84 (0.27, 2.64) 0.98 (0.26, 3.68) High school diploma or GED (Ref) 1.00 1.00 Some college, vocational training 1.52 (0.60, 3.83) 0.72 (0.22, 2.32) College degree 4.65 (0.87, 24.91) 3.09 (0.49, 19.66)

Maternal education

< high school 0.23* (0.06, 0.92) 0.28 (0.06, 1.31) High school (Ref) 1.00 1.00 Some/grad tech, or college 0.63 (0.22, 1.87) 0.60 (0.18, 2.03) > college grad 0.89 (0.31, 2.53) 0.66 (0.21, 2.14) Missing 0.60 (0.13, 2.83) 0.56 (0.08, 3.85)

Insurance status

None 1.06 (0.40, 2.83) 1.56 (0.48, 5.08) Medicaid 0.61 (0 .2 3 ,1 .6 3) 0.48 (0.14, 1.58) Private/other (Ref) 1.00

R e p r o d u c t i v e h i s t o r y

1.00

How difficult to make the decision

Very or somewhat difficult 0 .2 8 ** (0.13, 0.64) 0 .1 9 * * * (0.07, 0.49) Very or somewhat easy, not easy or difficult (Ref) 1.00 1.00

Facility gestational age limit 0 .7 4 * * * (0.68, 0.80) 0 .6 8 * * * (0.61, 0.77) Deviation from site gestational age 1.21 (0.85, 1.72) 0.99 (0.68, 1.43) Previous children

0 (Ref) 1.00 1.00 > 1 0.85 (0 .3 9 ,1 .8 5) 1.83 (0.65, 5.13)

Previous abortions (before index abortion)

0 (Ref) 1.00 1.00 > 1 1.49 (0.67, 3.29) 1.57 (0.60, 4.11)

Note. Cl – confidence interval; GED – general equivalency diploma; OR – odds ratio. The sample size was n = 226. * P < .0 5 ; * * P < .01; * * * P < .0 0 1 .

tum aw ays and near-limit abortion patients, and it is unlikely that the missing data biased the estimates. Third, our estimate o f more than 4 0 0 0 denied abortions was subject to som e amount o f error. It was based o n the proportion o f w om en turned away at 4 abortion facilities. W e assumed that these 4 facilities, which had com plete data on w om en turned away because o f gestational limits, w ere representative o f all

101 last stop facilities. At the sam e time, we expect the estimate is a conservative one because it did not include, for example, w om en w ho could not raise the funds to cover the procedure, take time off work, or get parental permission. It also did not in d u d e w om en w h o did not present for abortion care because they called ahead and realized they were b e yond the gestational limit. The total number of

wom en who did not obtain a desired abortion is likely much greater than 4 0 0 0 .

C o n c lu s io n s

W om en seeking abortions are more eco nomically disadvantaged than the larger popu lation o f w om en.33 W om en in need of second-trimester abortions are particularly vulnerable insofar as there are fewer providers that offer these services, and when they are available, procedures typically cost several hun dred, or even thousands, more dollars than a first-trimester procedure. Laws that impose lower and lower gestational limits will exacer bate the burdens these wom en face, and almost certainly, result in more unintended births.

About th e Authors Ushma D. Upadhyay, Tracy A. Weitz, Rana E. Barar, and Diana Greene Foster are with Advancing New Standards in Reproductive Health (ANS1RH), Bixby Center f o r Global Reproductive Health, and the Department o f Obstetrics, Gynecology and Reproductive Sciences, University o f California, San Francisco. Rachel K Jones is with the Guttmacher Institute, New York, NY.

Correspondence should be sent to Ushma D. Upadhyay, PhD, MPH, Advancing New Standards in Reproductive Health (ANSIRH), Bixby Center f o r Global Reproductive Health, University o f California, San Francisco, 1 3 3 0 Broadway, Suite 1100, Oakland, CA 9 4 6 1 2 (e-mail: [email protected]). Reprints can be ordered at http://www.aph.otg by clicking the Reprints” link.

This article was accepted April 5, 2 0 1 3 .

Contributors D. G. Foster was responsible for conceptualizing and designing th e Turnaw ay Study. U. D. U padhyay devel oped the analysis plan for this article, analyzed the data, and drafted the article. R. K. Jones was responsible for analyzing and interpreting data from th e G uttm acher Institutes Abortion Provider Census. T. A. Weitz, R. K. Jones, R. E. Barar, and D. G. Foster interpreted the data, reviewed drafts of th e article, and provided substantive com m ents on its content. R. E. B arar also provided m anagem ent support for the Turnaw ay Study.

Acknowledgments This study was supported by research an d institutional grants from the Wallace Alexander Gerbode Foundation, the David and Lucile Packard Foundation, the William and Flora Hewlett Foundation, and an anonymous foundation.

W e thank H eather Gould and Sandy Stonesifer for study coordination and m anagement; Janine Carpenter, Undine D am ey, Ivette Gomez, Selena Phipps, Claire Schreiber, and Danielle Sinkford for conducting inter views; Michaela Ferrari and Elisette W eiss for project support; and Jay Fraser and John N euhaus for statistical and database assistance.

Human Participant Protection Study protocol and procedures for th e Turnaw ay Study received institutional review bo ard approval from the

September 2014, Vol 104, No. 9 | American Journal of Public Health Upadhyay et al. | Peer Reviewed | Research and Practice | 1693

R E S E A R C H A N D P R A C T IC E

University of California, San Francisco Committee on Human Research. The Abortion Provider Census re ceived approval from the Guttmacher Institutes institu tional review board.

References 1. Pazol K, Zane SB, Parker WY, Hall LR, Berg C, Cook DA. Abortion surveillance-United States, 2008. MMWR Surveill Summ. 2 0 1 1;60(15):1-41.

2. Finer LB, Frohwirth LF, Dauphinee LA, Singh S, Moore AM. Timing of steps and reasons for delays in obtaining abortions in the United States. Contraception. 2006;74(4):334-344.

3. Drey EA, Foster DG, Jackson RA, Lee SJ, Cardenas LH, Damey PD. Risk factors associated with presenting for abortion in the second trimester. Obstet Gynecol. 2006; 107(1): 128-135.

4. Foster DG, Jackson RA, Cosby K, Weitz TA, Damey PD, Drey EA. Predictors of delay in each step leading to an abortion. Contraception. 2008;77(4):289-293.

5. Jones RK, Kooistra K. Abortion incidence and access to services in the United States, 2008. Perspect SexReprod Health. 201 l;4 3 (l):4 1 -5 0 .

6. Roe v. Wade. 410 US 113, 178 (1973).

7. Jones BS, Weitz TA. Legal barriers to second- trimester abortion provision and public health conse quences. Am J Public Health. 2009;99(4):623-630.

8. Guttmacher Institute. State policies on later abor tions. Available at: http://www.guttmacher.org/ statecenter/spibs/spib_PLTA.pdf. Accessed September 10, 2 0 1 2 . 9. Van Bebber SL, Phillips KA, Weitz TA, Gould H, Stewart F. Patient costs for medication abortion: results from a study of five clinical practices. Womens Health Issues. 2006;16(1):4-13. 10. Dobkin L, Gould H, Barar R, et al. Implementing a prospective study of women seeking abortion in the United States: the challenges of recruitm ent prior to informed consent. W orking Paper. Oakland, CA: Ad vancing New Standards in Reproductive Health; 2 0 1 2 .

11. Roberts SC, Avalos LA, Sinkford D, Foster DG. Alcohol, tobacco and drug use as reasons for abortion. Alcohol Alcohol. 2 0 1 2;47(6):640-648.

12. Foster DG, Kimport K, Gould H, Roberts SC, Weitz TA. Effect of abortion protesters on womens emotional response to abortion. Contraception. 2013;87(1): 8 1 -8 7 .

13. Henshaw SK. Abortion incidence and services in the United States, 1995-1996. Fam Plann Perspect. 1998;30 (6):263-270, 287.

14. Henshaw SK, Silverman J. The characteristics and prior contraceptive use of US abortion patients. Fam Plann Perspect. 1988;20(4): 158-168.

15. Henshaw SK, Van Vort J. Abortion services in the United States, 1987 and 1988. Fam Plann Perspect. 1990;22(3): 1 0 2-108, 142.

16. Jones RK, Zolna MR, Henshaw SK, Finer LB. Abortion in the United States: incidence and access to services, 2005. Perspect SexReprod Health. 2008;40(1): 6-16.

17. Ozimek A, Miles D. TRAVELTIME: Stata module to generate travel time and travel distance information. In: Statistical Software Components. College Station, TX:

StataCorp; 2012. Available at: http://econpapers.repec. org/software/bocbocode/s457449.htm. Accessed July 18, 2012.

18. Gipson JD, Koenig MA, Hindin MJ. The effects of unintended pregnancy on infant, child, and parental health: a review of the literature. Stud Fam Plann. 2 008;39(l):18-38.

19. Kost K, Landry DJ, Darroch JE. The effects of pregnancy planning status on birth outcomes and infant care. Fam Plann Perspect. 1998;30(5):223-230.

20. Joyce TJ, Kaestner R, Korenman S. The effect of pregnancy intention on child development. Demography. 2000;37(l):83-94.

21. Korenman S, Kaestner R, Joyce T. Consequences for infants of parental disagreement in pregnancy in tention. Perspect Sex Reprod Health. 2002;34(4):198- 205.

22. Monea E, Thomas A. Unintended pregnancy and taxpayer spending. Perspect Sex Reprod Health. 2011 ;43 (2):88-93.

23. Elfenbein DS, Felice ME. Adolescent pregnancy. Pediatr Clin North Am. 2003;50(4):781-800, viii.

24. Guttmacher Institute. State funding of abortion under Medicaid. Available at: http://w w w .guttmacher. org/statecenter/spibs/spib_SFAM.pdf. Accessed February 27, 2013.

2 5. Henshaw S, Joyce T, Dennis A, Finer L, Blanchard K. Restrictions on Medicaid funding for abortions: a litera ture review. Guttmacher Institute. 2009. Available at: http ://www.guttmacher.org/pubs/MedicaidLitReview. pdf. Accessed May 21, 2013.

26. National Abortion Federation. NAF hotline. Avail able at: http://www.prochoice.org/Pregnant/hotline/ index.html. Accessed July 18, 2012.

27. National Network of Abortion Funds. Get help. Available at: http://www.fundabortionnow.org/get-help. Accessed July 18, 2012.

28. Weitz T. The landscape of later abortions in the aftermath of the Tiller murder. Paper presented at: Abortion Discussion Group; Bixby Center for Global Reproductive Health; September 8,2 0 1 0 ; San Francisco, CA.

29. Joffe C. Dispatches from the Abortion Wars: The Costs o f Fanaticism to Doctors, Patients, and the Rest o f Us. Boston, MA: Beacon Press; 2010.

30. Joffe C. Doctors o f Conscience: The Struggle to Provide Abortion Before and After Roe v. Wade. Boston, MA: Beacon Press; 1996.

31. Harris LH, Debbink M, Martin L, Hassinger J. Dynamics of stigma in abortion work: findings from a pilot study of the Providers Share Workshop. Soc Sci Med. 2 0 1 1;73(7):1062-1070.

32. Freedman L, Landy U, Damey P, Steinauer J. Obstacles to the integration of abortion into obstetrics and gynecology practice. Perspect Sex Reprod Health. 2010;42(3): 146-151.

33. Jones RK, Finer LB, Singh S. Characteristics of US abortion patients, 2008. New York, NY: Guttmacher Institute. 2010. Available at: http://www.guttmacher. .pdf. Accessed May 21, 2013.

N E W RELEASE! Injury Prevention for Children and Adolescents, Second Edition Edited by Karen Liller, PhD

Injuries continue to be the leading cause o f death for children and adolescents in the United States. Because o f this trem endous public health burden, this second edition of Injury Prevention fo r Children and Adolescents: Research,

Practice, and Advocacy will continue to reach no t only academicians, practitioners, and students with the latest data and multifaceted intervention strategies, b u t also put a face on injury o f the innocent victims and their families. M ajor child a n d adolescent injuries are covered along with two new chapters that have great relevance for child and adolescent injuries. These chapters include Global Injuries Am ong C hildren and Adolescents a n d Translational Research Related to C hild and Adolescent Injury. The new edition will continue to provide a fram ew ork for practitioners, program plan ners, evaluators, and injury researchers that will serve the injury com m unity now and into the future.

Published in January 2012

OAPHA PRESS AN IMPRINT OF AMERICAN PUBLIC HEALTH ASSOCIATION

ORDER TODAY! 2 0 1 2 , 3 7 2 p ag e s , s o ftc o v e r ISBN 9 7 8 – 0 – 8 7 5 5 3 – 0 0 5 – 5

APHA Member Price: $49.00, L is t Price: $ 7 0 . 0 0

ORDER ONLINE: www.aphabookstore.org E-MAIL: A P H A @ pb d .co m

TEL: 8 8 8 – 3 2 0 – A P H A FAX: 8 8 8 -3 6 1 -A P H A

1 6 9 4 | Research and Practice | Peer Reviewed | Upadhyay e t al. American Journal o f Public Health | Septem ber 2 0 1 4 , Vol 10 4 , No. 9

Copyright of American Journal of Public Health is the property of American Public Health Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users may print, download, or email articles for individual use.