CONTEXTApproximately 45% of U.S. pregnancies each year are unintended, with higher
rates among women aged less than 25 years, some racial or ethnic minority groups,
and those with lower incomes.1
Unintended pregnancy is preventable with correct and continued contraceptive use.
Contraceptive counseling by trained healthcare professionals may help women, men,
and couples achieve their reproductive goals by supporting them to choose a method
concordant with their goals and preferences, and use their chosen method
correctly.
During 2010–2011, the Centers for Disease Control and Prevention and
the Office of Population Affairs conducted a series of systematic reviews, including
one summarizing the evidence on the impact of contraceptive counseling provided in
clinical settings on reproductive health outcomes.2 That review assessed evidence published from
January 1, 1985, to February 28, 2011, and identified promising components of
counseling (e.g., developing rapport, personalizing discussions to meet
clients’ individual needs, and addressing psychosocial determinants of
contraceptive use behaviors). Along with expert feedback and findings from two other
complementary systematic reviews on the impact of education and reminder systems in
family planning programs,3,4 the information was used to develop national
recommendations for providing quality family planning services.5 As part of a process to keep the
recommendations updated and revised as needed, the intent was to update the
systematic reviews (and consider new review topics) every 3–4
years.6
The objective of this systematic review is to update the prior review and
summarize the evidence on the impact of contraceptive counseling provided in
clinical settings. The newly identified evidence from the updated search and the
cumulative evidence from the initial and updated searches are summarized.
EVIDENCE ACQUISITIONThis systematic review is reported according to the PRISMA
checklist.7 The methods for
conducting this updated systematic review were similar to the approach used in the
prior reviews and have been described elsewhere.8 Briefly, as part of a series of systematic reviews on the
impact of contraceptive counseling and education in clinical settings, six key
questions (KQs) were developed (Appendix Table 1, available online). An analytic framework (Appendix Figure 1, available
online) was then applied to show the relationships between the population of
interest (women of reproductive age receiving Services in a clinical setting); the
intervention of interest (contraceptive counseling); and the outcomes of interest
Outcomes of interest relevant to KQs 1–3 included client experiences (e.g.,
satisfaction with the clinic visit, the provider, or counseling received);
short-term outcomes (e.g., increased knowledge); medium-term outcomes (e.g.,
increased contraceptive use); and long-term outcomes (e.g., decreased unintended
pregnancy). KQs 4–5 examined barriers and facilitators for clinics in
adopting and implementing contraceptive counseling in the family planning setting
and unintended negative consequences. This review describes the evidence for KQs
1–5; the evidence for KQ6 is described elsewhere.9
Search strategies were developed that included the identification of key
terms (Appendix Table 2,
available online), which were used to search multiple databases, including PubMed,
to identify potential articles published from March 1, 2011, to November 30,
2016.
Selection of StudiesRetrieval and inclusion criteria were developed a priori. Eligible
studies met the following criteria: conducted in the U.S., Canada, Australia,
New Zealand, or European countries categorized as “very high” on
the Human Development Index10;
published in English from March 1, 2011, to November 30, 2016; described a study
that addresses at least one KQ; and were full-length articles (abstracts and
letters to the editor were excluded). RCTs, nonrandomized trials, cohort, and
case-control studies were included. Articles also must have evaluated counseling
in a clinic-based setting. Contraceptive counseling was defined as an
interactive process between provider and client intended to help the client
achieve a reproductive health goal related to pregnancy prevention. Studies that
examined the presence or absence of contraceptive counseling in general, as well
as studies that examined specific counseling interventions (i.e., enhanced
models to standard of care) or approaches (e.g., personalization) were included.
Studies that focused solely on prevention of HIV or sexually transmitted
infections without a family planning component, or only included males, were not
considered.
Some inclusion criteria were specific to KQs. KQs 1–3 sought to
examine the relationships between contraceptive counseling and improved long-,
medium-, and short-term outcomes and client experiences; thus, included studies
had to have a comparison group. Among included studies, those that also examined
barriers and facilitators or unintended negative consequences met the inclusion
criteria for KQ 4 or 5. Articles that described a multi-component program (e.g.,
counseling in addition to a non-counseling component) had to report the impact
of the counseling component independent of the non-counseling component.
Data Abstraction, Assessment of Study Quality, and Synthesis of DataDetailed information (including study design, setting, population,
intervention, outcomes, and results) was abstracted by a team of four
abstractors and reviewed by two authors for relevance to contraceptive
counseling; differences were reconciled by consensus. The quality of each piece
of evidence was assessed using the grading system developed by the U.S.
Preventive Services Task Force, and risk of bias was rated as low, moderate, or
high.11 Several study
factors were considered when assessing quality, including design, recruitment
procedures, outcome measurement, adjustment for potential confounders,
attrition, and length of follow-up. Findings are reported separately for studies
conducted among adolescents and young adults (aged ≤ 24 years
[adolescents were defined as those aged ≤ 21 years; young adults were
defined as those aged 22–24 years]) and those conducted among adults or
mixed populations (adults and adolescents). This is because adolescents and
young adults are a special population of interest with potential unique needs.
Findings are also stratified by outcome type (long-, medium-, and short-term
outcomes and client experiences); as such, studies that examined multiple
outcomes may be discussed more than once. Summary measures of association were
not computed across studies because of the diversity of the interventions, study
designs, populations, and outcomes.
EVIDENCE SYNTHESISThe updated search strategy identified 24,953 articles (Figure 1). After removal of duplicates
(n=10,157) and applying the retrieval criteria, 205 full-text
articles were reviewed. Of these, ten studies (from 12 articles) met the inclusion
criteria.12–23 Findings from two articles among
adolescents are considered a single piece of evidence,17,18 as
are two articles from one study among adults and adolescents15,19;
only the most recent of these will be cited moving forward. Of the ten newly
identified studies included in this review, three examined the impact of
contraceptive counseling among adolescents and young adults,12,18,20 and seven examined the impact
among adults or mixed populations.13–16,21–23 Although most new studies were conducted among the general
population of women seeking contraceptive Services, one was conducted among a sample
of women post-abortion22 and two
were conducted among postpartum women.14,23
In combination with the initial review, the cumulative body of evidence
includes 32 studies that examined the impact of contraceptive counseling—nine
among adolescents and young adults and 23 among adults or mixed populations. Only
newly identified studies from the updated search are summarized in detail in the
text, although the cumulative evidence from both the initial and updated searches is
summarized in Table 1. For the evidence on
adolescents and young adults, Table 2 tallies
the significance of findings for behavioral outcomes and client experiences and
Appendix Table 3
(available online) describes details of each study. For the evidence on adults or
mixed populations, Table 3 tallies the
significance of findings for behavioral outcomes and client experiences and Appendix Table 4 (available
online) describes details of each study.
Adolescents and Young AdultsThree newly identified studies examined the impact of contraceptive
counseling among adolescents and young adults,12,18,20 each
examining a specific intervention compared with standard counseling. Two studies
were RCTs, one rated as having high risk for bias12 and the other rated as having moderate
risk for bias20; and one was a
pre—post study with a historical comparison group rated as having high
risk for bias.18 One study
examined a long-term outcome12
and all three studies examined at least one medium-term outcome.12,18,20 None of the
studies examined short-term outcomes, client experiences, barriers or
facilitators for clinics, or unintended negative consequences. Sample sizes in
the three studies ranged from 8718 to 1,155,12 and all participants were aged 14–24 years and
recruited from health clinics. A different model or approach to contraceptive
counseling was used in each of the three studies.
Long-term outcomes.One new study examined the impact of contraceptive counseling on
reducing pregnancy rates over 12 months and did not find a statistically
significant impact.12 In
this RCT by Berenson et al.12 evaluating a counseling intervention to improve
contraceptive adherence, 1,155 low-income adolescents and young adults aged
16–24 years requesting oral contraception (OC) were recruited from
five publicly funded reproductive health clinics. Participants were
randomized to receive either (1) face-to-face behavioral counseling that
utilized techniques based on the health belief model
(n=383); (2) this same counseling followed by monthly phone
calls for 6 months from a counselor (n=384); or (3)
standard care counseling (n=388). Pregnancy rates did not
differ between study groups over 12 months of follow-up.
Medium-term outcomes.All three newly identified studies examined at least one medium-term
outcome,12,18,20 with two finding at least one significantly
positive effect.12,20 The RCT by Berenson et
al.,12 described
above, found that condom use at last sexual intercourse did not differ by
study group at 3, 6, or 12 months. However, after adjustment, women who
received counseling plus monthly phone calls for 6 months were more likely
to report condom use at last sexual intercourse than those who received
standard counseling (OR=1.32, 95% CI=1.03, 1.70). The study also examined
but did not find a statistically significant impact of the counseling
interventions on OC adherence, OC continuation, or dual-method contraceptive
use (OC plus a condom) at 3, 6, or 12 months.12
The second new study examined the impact of contraceptive counseling
on consistent condom use.20 In this RCT, 828 adolescents aged 14–17 years
were recruited from five urban Title X—funded family planning
clinics. Adolescents were randomized to receive either a trans-theoretical
model—tailored counseling intervention designed to increase condom
use (n =424) or standard counseling
(n=404). Consistent condom use (defined as using condoms
during every sex occasion in the past month or past 3 months if no sex in
the past month) was significantly higher among adolescents in the
intervention versus control group at 6 months (61% vs 46%, respectively) and
12 months (51% vs 39%, respectively) but not 18 months
(p-values not reported).
The third new study examined the impact of contraceptive counseling
using if—then plans to overcome barriers to contraceptive use on
repeat visits for emergency contraception or pregnancy testing.18 In this pre—post
study from the United Kingdom with a historical comparison group, 87
adolescents aged 14–19 years using user-dependent contraceptive
methods (injections, pills, condoms) were recruited from a family planning
clinic. The number of consultations for emergency contraception and
pregnancy testing were compared from baseline (9 months before the
intervention) to follow-up (9 months after the intervention); intervention
participants were also compared with 131 teen girls recruited 1-year prior
who received standard counseling. Consultations for emergency contraception
or pregnancy testing decreased from baseline (49%) to follow-up (34%) among
the intervention group (statistical testing not reported), compared with a
nonsignificant reduction from 57% to 53% among the historical control
group.
Adults or Mixed Populations (Adults and Adolescents)Seven newly identified studies examined the impact of contraceptive
counseling among adults or mixed populations13–16,21–23: Two examined the impact of counseling,
in general, compared with no counseling,14,23 and five
examined a specific counseling intervention or aspect of counseling compared
with standard of care counseling.13,15,16,21,22 One study was
an RCT rated as having moderate risk for bias22; two were prospective cohort studies,
both rated as having moderate risk for bias15,16; one was a
retrospective cohort study rated as having high risk for bias21; and three were cohort
analyses of program or cross-sectional survey data, all rated as having high
risk for bias.13,14,23 All seven studies examined at least one medium-term
outcome,13–16,21–23 one
examined client experiences,22
and one study examined barriers or facilitators for clinics in adopting and
implementing contraceptive counseling interventions.22 None of the studies reported on
long-term outcomes, short-term outcomes, or unintended negative consequences.
Sample sizes in the seven studies ranged from 5021 to 193,310,14 and the age of participants ranged from
14 to 53 years. Participants were recruited from health clinics in four
studies,13,15,21,22 through
self-referral in one,16 and
from birth certificates in two studies.14,23
Medium-term outcomes.All seven newly identified studies examined at least one medium-term
outcome,13–16,21–23
with five finding a significantly positive effect on at least one
outcome.13–15,22,23
Two new studies examined contraceptive use, both of which found a
statistically significant impact of receiving counseling compared with not
receiving counseling.14,20 The first study by Cha and
colleagues14 was a
cohort analysis of cross-sectional survey data from 193,310 postpartum women
with a recent live birth, comparing women who reported receiving and not
receiving prenatal contraceptive counseling. Those who received prenatal
contraceptive counseling versus those who did not had higher odds of
postpartum contraceptive use (any method; 82% vs 72%, OR=1.72, 95% CI=1.64,
1.80). The second study by Zapata et al.23 was another cohort analysis of cross-sectional
survey data from postpartum women with a recent live birth reporting on
receipt of contraceptive counseling (n = 9,536); women who
received counseling both prenatally and postpartum and those who received
counseling during only one period were compared with those who did not
receive counseling during either period. Compared with postpartum women
receiving no counseling, those counseled during one time period (OR=2.01,
95% CI=1.55, 2.59) and both time periods (OR=2.74, 95% Cl=2.18, 3.45) had
higher odds of postpartum contraceptive use (69% vs 81% and 87%,
respectively; trend p-value<0.0001).
Five new studies of specific counseling interventions or aspects of
counseling examined use of more effective contraception,13,15,16,22,23 with four finding a statistically significant impact of
counseling.13,15,22,23 Whitaker
and colleagues22 reported
on one RCT conducted with 60 women aged 15–29 years presenting for
pregnancy termination at an urban academic clinic; participants either
received standard counseling (n=31), or Motivational
Interviewing—based counseling to encourage the use of highly
effective contraception (n=29). Significantly more women in
the intervention group than control group reported using a long-acting
reversible contraceptive (LARC) method at 1 (65.5% vs 32.3%, RR=2.03, 95%
CI=1.14, 3.61) and 3 months (60.0% vs 30.8%, RR=1.95, 95% 0=1.01, 3.77),
although more women who at baseline intended to use a LARC method
post-abortion were allocated to the intervention group. Use of any effective
method (an intrauterine device or a hormonal method) did not statistically
differ between women in the intervention and control groups at 1 or 3
months. In the second study, a prospective cohort study among 348 women aged
16–53 years seen for contraceptive care at six clinics by Dehlendorf
et al.,15 patients rated
their perception of the interpersonal quality of the counseling session
using a scale based on dimensions of patient-centered care. Counseling
sessions were also audio recorded, and research staff coded the
interpersonal communication behaviors of clinicians using a validated
scheme. Patients who reported high interpersonal quality of family planning
care versus those who did not were more likely to be using a highly or
moderately effective method at 6 months (66.0% vs 55.0%, OR=2.0, 95% 0=1.2,
3.5), although no audio recording—based measures of provider
communication behaviors were associated with use of a highly or moderately
effective method at 6 months. The third study was a cohort analysis of
program data from 771 women (mean age of 28 years) not seeking pregnancy and
receiving gynecological services from a system of Title X—funded
county primary care health centers. Electronic medical records were used to
compare women who did and did not receive (per provider report) reproductive
life plan—based counseling delivered by providers trained to have an
open dialogue with patients about their future life plans and the impact of
pregnancy and parenthood on these plans, and to discuss contraceptive
options in order of typical use efficacy.13 Results from multinomial logistic
regression suggested that receiving reproductive life plan—based
counseling was associated with LARC use versus no method or a non-medical
method (natural family planning or barrier method; OR=1.6, 95% 0=1.03,
2.61), but not associated with use of depot medroxyprogesterone acetate,
pills, patch or ring versus no method or a non-medical method. However, the
authors concluded no association based on results from model comparisons
(i.e., examination of the change in—2 log likelihood comparing models
with and without the reproductive life plan counseling variable). The fourth
study was a cohort analysis of cross-sectional survey data from postpartum
women by Zapata and colleagues2,5 described
above. In this study, compared with postpartum women who received no
counseling, the odds of using a more versus less effective contraceptive
method postpartum were higher for women counseled either prenatally or
postpartum (OR=2.10, 95% CI=1.65, 2.67) and for women counseled both
prenatally and postpartum (OR=2.33, 95% 0=1.87, 2.89, 32% vs 49% and 56%,
respectively; trend p-value<0.0001). The fifth
remaining study found no statistically significant effect of counseling on
LARC uptake after adjustment between women at intervention versus control
clinics as part of the Contraceptive CHOICE project, in which intervention
clinics used counseling where methods were presented in order of
effectiveness, although women in both groups had very high (>70%)
LARC uptake.16
Two new studies examined continuation of contraceptive use15,21; one of which found a statistically significant
positive impact of a specific aspect of counseling.15 In this prospective cohort study
described above by Dehlendorf et al.,15 patients were more likely to maintain contraceptive
method use at 6 months if they reported receiving high interpersonal quality
of family planning care (45.6% vs 36.1%, 0R=1.8, 95% CI=1.1, 3.0) and if
they had been seen by a provider rated as having high scores for
“invests in the beginning” (55.7% vs 36.8%, OR=2.3, 95%
CI=1.2, 4.3) and “elicits the patient perspective” (48.8% vs
38.1%, 0R=1.8, 95% CI=1.01, 3.2) based on audio-recorded counseling
sessions. However, neither “demonstrates empathy” nor
“invests in the end” were associated with contraceptive
continuation. Among a subanalysis of 67 adolescents aged 16—21 years,
patient—provider interactions around both method choice and method
use were coded as to whether they were interactive and appropriately
targeted. At the 6-month follow-up, use of a hormonal or LARC method among
adolescents was more common among those who had received interactive and
appropriately targeted counseling (80%) than among those who did not receive
this type of counseling (50%; no statistical testing was conducted). The
second study was a retrospective analysis of two cohorts, each with 25
women, receiving two different counseling approaches described as
“just try it” and “are you really sure.” This
study found no significant effect of one counseling approach over another on
implant continuation rates at 12 months.21
Client experiences.The RCT by Whitaker and colleagues22 assessed the effect of Motivational
Interviewing—based counseling on client experiences. During telephone
follow-up with participants at 3 months, significantly more women in the
intervention group (92%) compared with those in the control group (65%)
reported satisfaction with the counseling they received
(p=0.04).
Barriers and facilitators for clinics.The Whitaker and colleagues22 study provided the first evidence on barriers and
facilitators for clinics in adopting and implementing contraceptive
counseling interventions. The intervention lasted a median of 24 minutes
(range of 14–39 minutes) and the authors reported that no sessions
ended early, suggesting that clinic staff tolerated any disruption to clinic
flow because of the intervention. However, the authors also reported that
clinic staff did not approach 20% of eligible patients because of the
unavailability of a trained counselor after one of two trained counselors
left the institution. The training was described as resource intensive,
which prevented training a replacement. The authors commented that training
that is more efficient might be important for high-volume clinics that have
high staff turnover and many different staff members performing
counseling.
DISCUSSIONThe cumulative review included 32 studies published from January 1, 1985, to
November 30, 2016, that examined the impact of contraceptive counseling, in general,
or specific counseling interventions or aspects of counseling. Of nine studies
(three from the updated search12,18,20 and six from the initial review24–29) that examined the impact of counseling interventions among
adolescents and young adults, six found a statistically significant positive impact
of counseling on at least one outcome of interest.12,20,24,26,27,29 Twenty-three studies (seven from the updated
search13–16,21–23 and 16
from the initial review30–45) examined the impact of counseling
among adults or mixed populations, including five that examined counseling, in
general, compared with no counseling,14,23,39,44,45 and 18 that examined specific
counseling interventions or aspects.13,15,16,21,22,30–38,40–43 Of these, 16 found a statistically significant positive impact
of counseling on at least one outcome of interest.13–15,22,23,30–32,34–36,39,40,43–45
Overall, findings support the provision of contraceptive counseling,
compared with no counseling, on contraceptive use behaviors. With respect to
specific components of counseling, because of the diversity of interventions or
aspects of counseling examined—some of which incorporated multiple
approaches—this review was unable to assess the impact of a single counseling
component separate from the others or to compare the relative effectiveness of one
approach versus another. Nevertheless, components of counseling approaches that
resulted in some statistically significant positive change in outcomes of interest
can be considered when developing or updating contraceptive counseling guidelines
for family planning clients. Promising components to support method choice from the
cumulative body of evidence include an emphasis on the quality of interaction
between counselor and client (e.g., developing a rap-port)15,22,24,36,43,44; personalizing discussions to meet
clients’ individual needs15,20,22,29,34,35, 40,43; and addressing psychosocial determinants of contraceptive
use behaviors (e.g., perceived benefits and barriers, outcome
expectations).15,26,29,31,34,43 Other promising
components to support contraceptive behaviors once a method is chosen include
setting goals; discussing possible difficulties with achieving goals (e.g., side
effects, difficulty obtaining the method) and developing action plans to deal with
potential difficulties29,31,32,36; and multiple contacts with
clients20,24,26,27,35,36 New components
that emerged from the updated evidence that resulted in some statistically
significant positive effects include an emphasis on shared decision
making,15 asking about the
patient’s reproductive life plan/pregnancy intentions,13 and discussion of contraceptive methods by
level of effectiveness.13,22 Across the cumulative body of
evidence, seven of the counseling interventions that had a significant impact on
multiple outcomes15,22,24,26,32,35,40 included at least one of these promising
components.
LimitationsThis body of evidence has several limitations. Of the 32 studies in this
review, none were determined to have a low risk for bias, and half were
determined to have a high risk for bias.12–14,18,21,23,24,26–28,32,38–41,44,45 Studies were considered to be at risk for bias for
reasons including selection bias, recall bias, self-report bias, attrition bias,
or short follow-up times for behavioral outcomes. Among the RCTs examined,
primary weaknesses included lack of30,33,34,42 or no reporting of12,20,25,29,35,44 blinding; no reporting of allocation
procedures, including concealment12,25,29,30,35,37,43; and no concealment of allocation.34 Some studies were conducted among small
samples (fewer than 100 individuals)21,22,28,29,31,36,38,43,45 or stated not reaching sample size goals
based on power calculations20
and may have been underpowered to detect meaningful differences in outcomes.
Other studies did not state clearly how outcomes of interest were measured or
they used instruments with questionable validity.213,26,40 Some studies failed to adequately establish comparability
between study groups, limiting the ability to definitively attribute outcomes to
the counseling intervention.18,21,26,34,37,41,45 in one cohort
analysis of cross-sectional survey data, the temporal order between receipt of
contraceptive counseling during a recent clinical visit and contraceptive use at
last intercourse was uncertain. As previously mentioned, some studies were
conducted among subpopulations that may not represent the general population of
women seeking family planning Services, but represent groups that may especially
benefit from quality family planning services (i.e., postpartum women14,23,31,44 and those receiving post-pregnancy
termination care22,34,37,41–43 or services at a sexually
transmitted infection clinic).35 Further, limited new evidence was found that examined the
impact of contraceptive counseling on client experiences. Given the importance
of patient-centered care and the potential negative consequences of perceived
pressure to use specific contraceptive methods, understanding client perceptions
of patient-centered care and satisfaction are important. Also, studies examining
contraceptive counseling often do not consider client factors that might impede
contraceptive use behaviors, such as current intimate partner sexual or domestic
violence. Last, it is possible that additional articles meeting the inclusion
criteria for this systematic review have been published since the updated search
of the literature.
Despite these limitations, the evidence base for contraceptive
counseling also has several strengths. Nearly half of the studies (14 of 32) in
this review were RCTs,12,20,22,25,29–31,33,35,37,42,43 and many used random number tables or
computer randomization for group allocation,20,25,31,33,34,37,42,44 concealed
group allocation20,22,31,33,42,44 and used blinding.22,31,37,43 Several studies followed participants for at least 12
months.12,20,21,25,26,30–35 Other
strengths included high participation rates,15,21,29,32,33,41 high completion rates,33,35,41,44 small differences in follow-up rates
between study groups,20,30,32–34,37,42,44 and study
groups with similar baseline characteristics.12,20,25,29,30,32–36,42,44 Several
studies also used instruments with psychometric evidence of validity or
reliability to measure constructs of interest,15,26,36 and at least
three measured pregnancy by using urine tests versus self-report.24,32Several studies directly acknowledged training study
staff,16,18,22,24–29,38,42 and six
reported using standardized provider tools to improve intervention
implementation.12,15,20,22,28,30