Assignment: Counseling Clients Considering Abortion

Patient Education and Counseling 81 (2010) 362–367
Structured contraceptive counseling—A randomized controlled trial

Aileen M. Langston *, Linette Rosario, Carolyn L. Westhoff

Division of Family Planning and Preventive Services, Department of Obstetrics and Gynecology, Columbia University Medical Center, New York, USA

A R T I C L E I N F O

Article history:

Received 10 February 2010

Received in revised form 30 July 2010

Accepted 4 August 2010

Keywords:

Contraceptive counseling

Unintended pregnancy

Abortion

Contraception

Birth control

Structured counseling

A B S T R A C T

Objective: To evaluate the addition of structured contraceptive counseling to usual care on choice,

initiation, and continuation of very effective contraception after uterine aspiration.

Methods: We conducted a RCT of a version of the WHO Decision-Making Tool for Family Planning Clients

and Providers with women having a procedure for a spontaneous or induced abortion. Our intervention

provided structured, standardized counseling. We randomized women to usual care or usual care with

structured counseling. Our outcomes included choosing a very effective contraceptive method and 3

months continuation.

Results: Fifty-four percent of all participants chose a very effective method. Women in the intervention

group were no more likely to choose a very effective method (OR 0.74, 95% CI 0.44, 1.26) or to initiate

their method compared to the usual care group (OR 0.65, 95% CI 0.31, 1.34). In multivariate models,

structured counseling was not associated with using a very effective method at 3 months (AOR 1.06, 95%

CI 0.53, 2.14).

Conclusion: In this setting, structured counseling had little impact on contraceptive method choice,

initiation, or continuation.

Practice implications: Adding structured counseling did not increase the proportion choosing or

initiating very effective contraception in a practice setting where physicians already provide

individualized counseling.

� 2010 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at ScienceDirect

Patient Education and Counseling

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / p a t e d u c o u
1. Introduction

Despite the availability of very effective contraceptive methods,
the rate of unintended pregnancy in the United States (US)
remained stagnant at 49% from 1995 to 2001 [1]. Most women in
the US have not used the most effective methods available [2], and
47% have had a repeat abortion [3]. Women’s health professionals
have regarded counseling as an important component of improv-
ing contraceptive use [4], and access to counseling services has
been considered an integral part of informed choice [5]. The World
Health Organization (WHO) has supported the practice of
contraceptive counseling so that patients can make informed
decisions in conjunction with their provider [6].

Accepted practice within contemporary healthcare has been to
offer patients information regarding diagnoses and proposed
treatment options. Contraceptive counseling, where options are
presented with mechanisms of action, efficacy, risks and benefits,
has been a challenge due to the limited resources in the clinical
* Corresponding author at: Division of Family Planning and Preventive Services,

Columbia University Medical Center, 622 West 168th Street, PH 1669, New York,

NY 10032, USA. Tel.: +1 212 305 4805; fax: +1 212 305 6438.

E-mail address: [email protected] (A.M. Langston).

0738-3991/$ – see front matter � 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.pec.2010.08.006
setting and the ability of any person to receive and comprehend a
large amount of information.

A recent Cochrane Review found that ‘‘little evidence from
randomized controlled trials supports the hypothesis that
counseling improves contraceptive use,’’ and a systematic review
of the literature on counseling to prevent unintended pregnancy
also [7] found limited evidence regarding its effectiveness [8]. In
1996 the US Preventive Services Task Force (USPSTF) recom-
mended contraceptive counseling, but the 2002 USPSTF withdrew
this recommendation due to insufficient evidence [9,10].

Limited data has suggested a possible benefit to using
structured counseling—consisting of audio–visual materials with
standardized information—for contraceptive counseling. Two
randomized controlled trials (RCTs) utilized structured audio–
visual educational material with standardized information about
contraceptive methods. The results from both studies showed
increased contraceptive use or continuation of effective contracep-
tive methods (pill and injection) 1 year later [11,12].

In a post-abortion population, one RCT of counseling performed
by a contraceptive specialist along with advanced provision of
contraceptive methods compared to routine counseling found
increased uptake of long acting reversible contraceptives and
increased continuation at 4 months but no difference in repeat
abortion rates at 2 years [13]. The information given by the

http://dx.doi.org/10.1016/j.pec.2010.08.006

mailto:[email protected]

http://www.sciencedirect.com/science/journal/07383991

http://dx.doi.org/10.1016/j.pec.2010.08.006

A.M. Langston et al. / Patient Education and Counseling 81 (2010) 362–367 363
specialist counselor in this study was not standardized or given in a
structured audio-visual format.

In an attempt to meet family planning counseling needs, the
WHO developed a series of family planning guidelines and tools,
including the Decision-Making Tool for Family Planning Clients
and Providers (DMT) [14]. A double-sided flipchart with one side
for the client to aid in decision-making and the other side for the
provider to aid in the counseling process by giving information and
guidance, this tool was studied for improving communication with
clients in limited resource settings [15,16]. The DMT was found to
improve communication, particularly with clients choosing a new
contraceptive method [15].

Given the need for resource efficiency in health care, the belief
by many providers and organizations that contraceptive counsel-
ing is necessary and worthwhile, and the limited literature [7–
10,17], we aimed to study this topic using a structured and
standardized counseling intervention based on the DMT in a post-
abortion setting with most modern methods available for
immediate initiation. Our study evaluated structured, standard-
ized contraceptive counseling for its influence on participants
choosing a very effective contraceptive method at the time of first
trimester vacuum aspiration, method initiation, and 3 months and
6 months method continuation.

2. Methods

2.1. Setting and participants

From December 2008 to July 2009, we enrolled participants
from a family planning referral clinic to a private practice setting
serving a predominantly Hispanic (Dominican) population with
Medicaid coverage in New York City. Providers at the practice were
all physicians: faculty, fellows, and residents at Columbia
University Medical Center (CUMC). Vacuum aspiration procedures
were offered 1 day per week, and on a given day, 3–4 physicians
each cared for 6–8 patients.

The study population consisted of women seeking a first trimester
procedure for a spontaneous or induced abortion. Inclusion criteria
were (1) age � 18 years, (2) no desire to become pregnant right away,
(3) fluency in Spanish or English, and (4) access to a telephone. The
CUMC Institutional Review Board approved this study.

2.2. Structured contraceptive counseling intervention

In this study we sought to address whether structured,
standardized, non-directive counseling (the intervention) in the
setting where contraceptive methods are immediately available and
the women have confirmed fertility, will result in increased choosing
of very effective contraceptive methods, method initiation, and
method continuation at 3 months. Structured counseling consisted
of the trained research coordinator reading and displaying a
contraceptive flipchart in its entirety to the participant in a private
office with samples of each method available for patients to see and
touch. The counseling was structured in that the format included
visual and audio components allowing the participant to both
visualize and hear the information. The counseling was standardized
in that the same information was presented every time the
counseling was performed. Participants were encouraged to ask
questions and to write down questions for their physician on
supplied note cards. The research assistants were trained to answer
questions using only the information from the flipchart. If a question
was not able to be answered by the information on the flipchart, the
research assistant was instructed to request the participant ask her
provider this question during usual care.

The flipchart was a version of the WHO 2005 Decision-Making
Tool for Family Planning Clients and Providers (DMT) [14]. We did
not intend the intervention to provide tailored counseling, though
that is one of the common uses of the DMT. We chose to use the
format of the DMT for its ready-made structure, simple language
and images to create our structured and standardized intervention.
We utilized the portion of the DMT focused on choosing a method
and the method tabs (overview and information for choice,
medical eligibility criteria, possible side effects, how to use, when
to start, and what to remember). We modified this portion of the
DMT to add methods available in the US (patch, ring, levonorges-
trel IUD, and etonogestrel implant) and to remove information
about methods not available in the US (NET-EN injections, monthly
injections, and Norplant) or not appropriate for post procedure
patients (lactational amenorrhea and fertility awareness). Thus the
flipchart administered by a research coordinator gave our
intervention a structured format with visual and audio compo-
nents. The research coordinators gave standardized information
using this structured tool. The simple language and images in the
flipchart, as well as reading the pages aloud as they were viewed,
mitigated any effects of low literacy. Information on contraceptive
methods (female sterilization, male sterilization, copper IUD,
levonorgestrel IUD, etonogestrel implant, depo provera injection,
ring, patch, pill, and condom) was presented on 5–7 double-sided
pages per method with patient and counselor focused content on
the front and back, respectively. We used flipcharts printed in
English for our participants who preferred English and flipcharts
translated and printed in Spanish for our participants who
preferred Spanish. The flipchart included information on con-
traceptive effectiveness, how to use each method, possible side
effects, and when to seek help.

Usual care consisted of a single physician performing the
medical history, physical exam, ultrasound, obtaining informed
consent for the suction aspiration procedure, and carrying out this
procedure for each patient. This visit required about one hour to
complete. Contraceptive counseling was routinely offered by the
physician as well and was embedded in the visit. As part of usual
care, the content and duration of contraceptive counseling
performed by the provider was left to their discretion.

2.3. Study procedures

Two research coordinators fluent in English and Spanish
performed enrollment and follow-up. We used training scripts
and role play to standardize interactions with participants.
Questionnaires were piloted and adjusted based on responses
prior to enrollment.

We assessed all women aged 18 or older registered in the clinic
for eligibility. To ease anxiety, the coordinator first gave each
patient basic information about routine clinic procedures before
discussing the study. Interested and eligible women were
consented. A baseline questionnaire was administered to collect
demographic characteristics as well as partnership, reproductive,
and contraceptive histories. Participants were then randomized to
usual care with intervention versus usual care alone. Those
randomized to the intervention group received structured
counseling by a coordinator immediately prior to usual care
during the same visit. Attention was paid to minimize delay for
women in the intervention group.

Using a random-number table, we determined the sequence for
1:1 allocation constrained by blocks of 10. Randomization
assignments were sealed inside numbered, opaque envelopes.
The coordinator opened the next sequentially numbered envelope
after completing informed consent. No blinding of participants or
coordinators was feasible due to the nature of the intervention.
Physician-providers did not know the participant’s allocation
group, did not discuss the study with patients, and were asked not
to change their counseling.

A.M. Langston et al. / Patient Education and Counseling 81 (2010) 362–367364
Contraceptive methods available to participants immediately
following their procedure included intrauterine devices (IUDs),
implants, injections, rings, and pills. The IUDs and implants were
donated and available at no cost to all clinic patients. All
participants had either New York State Medicaid coverage for
prescription contraceptives or access to additional free supplies at
a safety net clinic so all contraceptives offered were available free
of charge. The patch was available by prescription only and
sterilization by referral only. Those who chose pill or ring received
either a prescription or a 1-month supply and prescription. All
participants received condoms with handouts on emergency
contraception and condom use.

After each enrollment day, we reviewed charts to confirm that a
procedure was performed and to identify the contraceptive
method chosen as well as whether initiation was immediate or
delayed. Coordinators called participants 3 months after enroll-
ment to assess contraceptive use. A subset of patients received 6
months follow-up phone calls. Initial analysis of the first 101
participants to complete both 3 and 6 months data found no
significant differences, so 6 months follow-up was stopped to focus
on maximizing 3 months follow-up.

2.4. Outcomes and analysis

The primary outcome of this study was proportion of
participants choosing a very effective contraceptive method.
Secondary outcomes were method initiation on the day of the
procedure and method continuation of very effective and/or
effective methods at 3 months, and at 6 months for the sub-group
for whom we collected data.

The WHO defined very effective contraceptive methods as those
with 1 year typical use pregnancy rates of <1% (sterilization, IUDs, and implants) [6]. Effective methods have typical use pregnancy rates of 1–9% (pills, rings, patches, and injections). The WHO defined additional categories for methods with �10% and >25%
typical use pregnancy rates. In this study, we used the WHO

[(Fig._1)TD$FIG]

Fig. 1. Participant enrollment and follow-up in a randomized con
definition for very effective and effective methods and defined less
effective methods as those with �10% pregnancy rate (condoms,
withdrawal, periodic abstinence, and no method).

We defined initiation of effective and very effective methods as
leaving the clinic with a method requiring no healthcare provider
contact to begin use. If a participant requested pills and left clinic
with a pill pack and a prescription, this was coded as immediate
initiation. If she left with a prescription only, this was considered
delayed initiation because she needed to go to a pharmacy to begin
using the method. Less effective contraceptive methods (condoms,
withdrawal, and periodic abstinence) were coitally dependent and,
therefore, were not able to be initiated in the clinic.

We defined continuation as using a contraceptive method at 3
or 6 months that was in the same effectiveness group as the
method requested at enrollment. For example, two patients
requested sterilization and were using an IUD at the 3 months
follow-up interview. Both these participants were counted as
‘continuers’ for the very effective group.

In this clinic in 2003–2004, 29% of patients chose the most
effective available methods (injection, copper IUD, or sterilization)
following a first trimester aspiration procedure [18]. We designed
our study to identify an increase from 30% to 50% of women
requesting a very effective method in the intervention arm. With
up to 20% loss due to exclusion after randomization, a two-sided
alpha of .05, and power of .80, we needed 125 women in each arm.

We used SAS, version 9.2 (SAS Institute, Cary, NC) for statistical
analyses to compare the intervention and control groups. We
performed Chi-square analyses to assess differences between
allocation groups. We calculated two-sided p-values and 95%
confidence intervals. We performed logistic regression analyses
with two dependent outcomes: very effective method use at 3
months; or very effective or effective method use at 3 months. The
8 participants who reported sexual abstinence since enrollment
due to no partner were excluded from these analyses. We
performed univariate logistic regression with (1) intervention,
(2) immediate initiation, (3) age, (4) education, (5) ethnicity, (6)
trolled trial of structured contraceptive counseling, NY 2009.

Table 2
Structured contraceptive counseling versus usual care: contraceptive method

chosen and 3 months continuation.

Intervention

(N = 114)

Usual care

(N = 108)

Total

(N = 222)

p-Value*

Contraceptive method chosen

Very effective methodsa 57 (50%) 62 (58%) 119 (54%) 0.27

Effective methodsa 48 (42%) 37 (34%) 85 (38%) 0.27

Less effective methodsa 9 (8%) 9 (8%) 18 (8%) 1.0

(N = 89) (N = 83) (N = 172b)

3 months continuation

Very effective methodsc 41/48 (85%) 40/52 (77%) 81/100 (81%) 0.28

Effective methodsc 28/41 (68%) 21/31 (68%) 49/72 (68%) 0.96

* Chi-square p-value.
a Very effective methods—copper IUD, levonorgestrel IUD, etonogestrel implant,

sterilization. Effective methods—DMPA, ring, patch, pill. Less effective methods—

intervention group: 1 undecided, 2 abstinence, 2 declined contraception, 4

condoms. Control group: 3 undecided, 1 natural family planning, 1 coitus

interruptus, 1 declined contraception, 3 condoms.
b Less effective methods are not represented in this total.
c Numerators are continuers and denominators are those who chose this method

group and completed 3 months follow-up.

A.M. Langston et al. / Patient Education and Counseling 81 (2010) 362–367 365
parity, (7) prior abortion, (8) stable relationship status, (9) provider
and (10) current smoking. Variables were chosen for multivariate
logistic regression based upon univariate results and overall
importance to the clinical outcome. We constructed the final
model using the likelihood ratio test as variables were added
sequentially to determine the most parsimonious model. The
Hosmer–Lemeshow statistic was calculated to test the goodness-
of-fit of the final model.

3. Results

3.1. Enrollment

We screened 380 women and enrolled 250 women (Fig. 1). We
excluded 28 women after randomization because they did not
have a procedure that day primarily due to pregnancies in the
second trimester, completed spontaneous abortions, and ectopic
pregnancies. The remaining 222 women were eligible for analysis
and follow-up.

The groups were well balanced with regard to baseline
characteristics (Table 1). They were mainly Hispanic and in stable
relationships, defined as a relationship the participant reported
will continue for >1 year. Participants’ ages ranged from 18 to 45
years with a mean age of 26.2 years. Participants were seeking
induced abortion (94%) or spontaneous abortion management
(6%). The time used to conduct the structured counseling
intervention was, on average, 20 min (standard deviation �8 min).

3.2. Methods requested

The intervention and control groups were similar in the
methods requested (Table 2). Participants in the intervention
group were similar to the usual care group in often choosing a very
effective method (OR 0.74, 95% CI 0.44, 1.26). Most women
requested very effective methods (levonorgestrel IUD (27%),
copper IUD (15%), implant (9%) and sterilization (2%)). Many
women requested effective methods (oral contraceptive pills
(18%), vaginal ring (9%), injection (7%), and patch (5%)). Fewer
women requested less effective methods (undecided/declined
(n = 7), condoms (n = 7), abstinence (n = 2), withdrawal (n = 1) and
periodic abstinence (n = 1)).

Comparing the demographics of participants who chose very
effective methods to those who did not, parous women and women
in a stable relationship were more likely to choose a very effective
method (OR 2.51, 95% CI 1.35, 4.67 and OR 1.98, 95% CI 1.11, 3.54,
Table 1
Demographics and reproductive history—structured contraceptive counseling

versus usual care (N = 222).

Intervention (N = 114)

N (%)

Usual care (N = 108)

N (%)

Age (SD) 25.6 (5.7) 26.8 (6.7)

Age < 25 years 59 (52%) 49 (45%) Hispanica 98 (87%) 97 (90%) Education�12th grade 74 (65%) 77 (71%) Birthplace United States 47 (41%) 43 (40%) Dominican Republic 50 (44%) 53 (49%) Other 17 (15%) 12 (11%) Current smokersb 20 (18%) 20 (19%) Gravida > 1 98 (86%) 96 (89%)

Parous 84 (74%) 81 (75%)

Ever had a prior abortion 57 (50%) 58 (54%)

Ever used contraception 109 (96%) 104 (96%)

Current stable relationship 78 (68%) 77 (71%)

a One missing value from intervention group.
b One missing value from intervention group.
respectively). Six providers saw the majority (91%) of the
participants. No differences were seen in the methods requested
(p = 0.44) or the proportion of methods initiated immediately
(p = 0.83) among these providers. There was no difference between
the intervention and control groups in the physician-providers
from whom they received usual care (p = .59).

3.3. Immediate versus delayed initiation

Participants in the intervention group were not more likely to
initiate the requested method immediately compared to those in
the usual care group (OR 0.65, 95% CI 0.31, 1.34) (Table 2). Only 15
percent of participants chose a method that could not be initiated
the same day (18 less effective methods, 10 patches, and 5
sterilizations). The other 189 participants selected a method that
could be initiated the same day; 80% of these women initiated their
method the same day (80 IUDs, 28 pills, 19 implants, 15 injections,
and 10 rings). The remaining 20% of participants had delayed
initiation of their method (14 IUDs, 11 pills, 10 rings, 1 injection,
and 1 implant). Of these, 3 women preferred to delay IUD insertion,
and 1 woman wanted to obtain her pill prescription from her
personal physician. The physician-providers delayed 9 initiations
due to infection, 2 due to bleeding, and 1 due to lack of
confirmatory products of conception at the time of the procedure.
Twenty-one women were given prescriptions only (10 rings, and
11 pills).

3.4. Follow-up and continuation

Of 222 participants, 186 (84%) completed 3 months follow-up
(Fig. 1). Loss to follow-up was equal between the intervention and
control group. The baseline characteristics and requested methods
of the women who completed 3 months follow-up and those who
did not were similar (data not shown). Those in the intervention
group who completed 3 months follow-up had chosen similar
methods compared to those in the control group (p = 0.51). No
participants reported a repeat pregnancy at 3 months.

For those who chose a very effective or effective method, 3
months continuation of the requested method and 3 months
continuation of immediately initiated methods were not signifi-
cantly different comparing the intervention group to the usual care
group (OR 1.24, 95% CI 0.62, 2.50 and OR 1.43, 95% CI 0.58, 3.52,
respectively) (Table 2). Fourteen (78%) participants who chose a
less effective method completed 3 months follow-up; 13 reported
being sexually active; and only 2 reported adopting an effective

Table 3B
Predictors of contraceptive method use at 3 months, multivariate model (N = 186).

Very effective method

use

Very effective or ef-

fective method use

AOR 95% CI AOR 95% CI

Structured counseling 1.06 (0.53, 2.14) 1.59 (0.77, 3.28)

Immediate initiation 15.5 (6.02, 39.7) 4.26 (2.05, 8.87)

Age 0.91 (0.43, 1.89) 1.67 (0.81, 3.47)

Education – – – –

Prior abortion – – – –

Parous 3.17 (1.37, 7.32) – –

Hispanic – – – –

Relationship – – – –

Smoking – – – –

*Excluded participants abstinent since enrollment from the analysis (N = 8). **The

reference group is ‘no’ and the comparison group is ‘yes’ except for age where

reference group is <25 years and comparison group is �25 years. A.M. Langston et al. / Patient Education and Counseling 81 (2010) 362–367366 method (pills). In a sub-group analysis of those who initiated a very effective method on the day of enrollment (n = 83), the interven- tion group trended towards increased 3 months continuation compared to the usual care group (98% versus 83%; p = .06). With the initial participants at the start of the study, we took the opportunity to conduct 6 months follow-up interviews. We completed 6 months follow-up with 131 (59%) participants. For these participants, 6 months continuation between the inter- vention group (67%) and the usual care group (68%) was similar (OR 0.95, 95% CI 0.45, 2.02). Two participants reported a repeat pregnancy at 6 months, one from each randomization group. 3.5. Predictors of using a very effective and/or effective method at 3 months When limiting our outcome to using a very effective method at 3 months, the counseling intervention did not have a strong effect in univariate or multivariate models (Tables 3A and 3B). In univariate analyses, completing at least the 12th grade in school and immediate initiation of a requested contraceptive method had the strongest associations with using a very effective or effective method at 3 months (Table 3A). In a multivariate model, the counseling intervention did not have a strong association with using a very effective or effective method at 3 months (AOR 1.59, 95% CI 0.77, 3.28). 4. Discussion and conclusion 4.1. Discussion We sought to address whether structured, standardized, non- directive counseling (the intervention) in the setting where contraceptive methods are immediately available and the women have confirmed fertility, will result in increased choosing of very effective contraceptive methods. We specifically chose a counsel- ing format that would not be performed by a physician to reflect the reality of limited health resources and the common practice of family planning clinics in the US to utilize non-physicians to perform counseling. We chose standardized counseling in contrast to tailored counseling to ensure that participants in the interven- tion group received the same information to minimize bias from the counselor. Minority women have been shown to be more likely to receive contraceptive and sterilization counseling compared to white women [19], and our clinic serves a predominantly minority population. We performed a RCT of an intervention utilizing a modified version of a readily reproducible counseling intervention (DMT) Table 3A Predictors of method use at 3 months, univariate analyses (N = 186). Very effective method use Very effective or effective method use OR 95% CI OR 95% CI Structured counseling 0.97 (0.53, 1.74) 1.35 (0.68, 2.68) Immediate initiation 14.02 (5.58, 35.22) 3.87 (1.90, 7.89) Age 1.05 (0.58, 1.89) 1.44 (0.73, 2.86) Education 1.65 (0.87, 3.14) 2.11 (1.04, 4.25) Prior abortion 0.91 (0.50, 1.64) 1.62 (0.82, 3.22) Parous 2.43 (1.19, 4.95) 1.37 (0.65, 2.89) Hispanic 0.86 (0.33, 2.24) 1.46 (0.52, 4.10) Relationship 1.38 (0.72, 2.65) 1.10 (0.53, 2.29) Smoking 0.77 (0.35, 1.69) 1.12 (0.46, 2.74) *Excluded participants abstinent since enrollment from the analysis (N = 8). **The reference group is ‘no’ and the comparison group is ‘yes’ except for age where reference group is <25 years and comparison group is �25 years. that is available online and developed by experts. Our structured and standardized counseling intervention did not result in more women choosing a very effective contraceptive method, immedi- ately initiating more methods, or significantly increasing 3 months continuation of their chosen method in our setting. In our clinic, physicians who specialize in family planning are providing contraceptive counseling with the patients as an integrated part of their visit for a first trimester uterine aspiration. Additional counseling may have been unnecessary in this setting. Our study had several limitations. Our clinic setting had specialized providers as well as a specific ethnic demographic that limited the generalizability of our study’s findings. We utilized the DMT for structured, standardized counseling, and it was designed for tailored counseling. This approach may have affected the effectiveness of the intervention. Our 3 months contraceptive data was self-reported and vulnerable to social desirability bias. We made an effort to reduce patients’ anxiety but could not eliminate it before the intervention. This anxiety could have lessened the effects of the structured counseling intervention. A further limitation was that the providers in our setting were aware of the study and could have altered their …