This review complements findings of the accompanying review on contraceptive counseling,²⁰ defined as an interactive process between a provider and client intended to help the client achieve a reproductive health goal. This education review focuses more narrowly on the process of helping clients increase their knowledge of contraception so they can make informed decisions about using contraception, and then correctly and consistently use their chosen method.

This systematic review is reported according to the PRISMA checklist.²¹ The approach used was similar to that implemented in previous reviews²² guided by an analytic framework showing the logical relationships between the population of interest, the interventions, and the outcomes of interest (Figure 1). This review of education interventions primarily focused on short-term outcomes related to contraceptive knowledge and decision making; studies were excluded if the intervention they described focused only on skills or behaviors. Of seven key questions (KQs; Table 1), the first five (KQ1–KQ5) asked whether educational interventions affected the following: contraceptive knowledge, including correct method use, risks and benefits, side effects, and method effectiveness; participation in decision making; satisfaction/comfort with services and decision making; attitudes toward contraception; and for interventions providing education on the full range of methods, selection of more versus less effective methods (KQ5). KQ6 was expanded from a focus in the original review on health literacy and asked if there were any barriers or facilitators for clinics/clients; KQ7 was added to address unintended consequences.

The search strategy included terms common to the other systematic reviews in this series on contraceptive counseling, reminder systems, and patient preferences for contraceptive counseling. Articles were identified from 16 electronic databases (Appendix Table 1, available online), and supplemented by hand searches of the bibliographies of identified articles. The initial and updated reviews identified articles published between January 1985–February 2011 and March 2011–November 2017, respectively.

Retrieval and inclusion criteria were developed a priori. Included studies were conducted in a clinic-based setting where family planning services: were provided or assessed an intervention that could be implemented feasibly in a clinic-based setting (e.g., not a multisession course series); contained a comparison group; were conducted in the U.S., Canada, Australia, New Zealand, or a European country categorized as “very high” on the Human Development Index²³; and addressed at least one KQ. Studies focusing primarily on sexually transmitted infections were included only if they incorporated education on condom use to prevent pregnancy. For studies meeting inclusion criteria, outcomes of interest to other systematic reviews in this series (i.e., behavioral outcomes, such as contraceptive use, and long-term outcomes, such as pregnancy) also were evaluated because of their potential to be mediated by knowledge. Outcomes related to the KQs were considered primary outcomes; all other outcomes were considered secondary outcomes.

Detailed information was abstracted by a team of four abstractors and reviewed by two authors for relevance to contraceptive education; differences were reconciled by consensus. In provider-enhanced interventions, a healthcare provider/educator interacted with the intervention to help participants in at least one group understand or use the information presented. By contrast, in provider-independent interventions a healthcare provider/educator did not go beyond the standard of care to facilitate participant understanding; the participant received the intervention either before or after, rather than during their appointment, and providers were given no specific instructions to help participants understand or make use of information. Studies were classified as evaluating more than one intervention if they included distinct mediums (e.g., written materials versus audio/video) or a provider-independent and a provider-enhanced intervention, each of which was separately compared with a control arm.

Study quality was assessed using a modification of the grading system developed by the U.S. Preventive Services Task Force.²⁴ Studies were first rated based on the U.S. Preventive Services Task Force evidence scale and then specific criteria were used to judge whether the study had high, moderate, or low risk for bias.

Data synthesis was primarily qualitative. Meta-analysis was not performed because of the large degree of heterogeneity across studies with respect to study design, study populations, lengths of follow-up, and outcomes measured.

Evidence from the initial and updated systematic reviews (Figure 2), including primary and secondary outcomes, is summarized in Figure 3 and presented in detail in Appendix Table 2 (available online). Primary outcomes from the review update are further discussed in the text.

The systematic review identified 14,796 unique abstracts. From the title and abstract review, 14,591 articles were excluded and 205 were retrieved; 17 met inclusion criteria.^25–41 However, because three studies including the same population and intervention were described in multiple articles,^{25,29,30,32,34,39} these articles were counted as 14 independent studies. Hand searches of references from included articles identified one additional article⁴² providing further details on a study from the review update,^25,32 and one article,⁴³ providing further details on a study from the original review.⁴⁴ Together with evidence from the original review,¹⁸ 31 studies described in 37 articles are included in this review update (Figure 2).^25–61

Of the included studies, 17 were RCTs, of which nine were classified as having low,^{27,29,30,33,43,44,47,49,52,53,58,60} six moderate,^{25,26,31,32,35,40,42,46} and two high^38,51 risk for bias. Three were non-RCTs with separate groups of intervention and control participants; of these, two were classified as having moderate^45,54 and one high⁵⁷ risk for bias. Nine used a pre-/post-test study design, of which three were classified as having moderate^36,37,41 and six high^{28,50,55,56,59,61} risk for bias. Two final studies, also with high risk for bias, used a sequential cohort analysis in which one set of participants was studied prior to implementation of an intervention and was compared with another set of participants studied during the intervention period.^34,39,48

Of the 31 studies identified, five included two interventions (one included a written intervention and an audio/video intervention⁴⁶; four included a provider-independent and a provider-enhanced intervention^{29,30,43,44,47,52,53}) that were each separately compared with a control arm, for a total of 36 interventions. In addition to being assessed for their effects on the a priori outcomes of interest (Figure 3), for three interventions, knowledge gains were evaluated for their further effect on these outcomes, including positive attitudes toward oral contraceptives (OCs)⁴¹ and long-acting reversible contraceptives (LARC),³³ intention to use OCs,⁴¹ and continued use of OCs.^25,42

The review update identified 12 interventions^{25,27,28,31–42} assessed for their effect on knowledge (KQ1), including knowledge of correct method use, risks and benefits, side effects, and method effectiveness; of these, ten had a significant positive effect.^{25,27,28,31–37,39,41,42} Details are described separately for provider-independent and provider-enhanced interventions.

Four interventions in the review update were assessed for their effect on satisfaction/comfort with services and decision making (KQ3): three were provider-independent interventions assessed through an RCT,^26,33,40 and one was a provider-enhanced intervention assessed through a sequential cohort analysis.^34,39 Of the provider-independent interventions, the first³³ found that intervention participants given the opportunity to interact with a Facebook page, versus control participants allowed to view pamphlets with the same information, had higher composite scores for satisfaction with the counseling they received (p<0.001). By contrast, the second intervention⁴⁰ found a lower percentage of intervention participants using a smartphone app, versus control participants receiving standard-of-care counseling from a health educator, who reported they were very satisfied with the counseling they received (p<0.001). The third provider-independent intervention²⁶ was a video informed by the transtheoretical model of behavioral change designed to increase awareness and dispel misconceptions about LARC. Intervention and control participants did not differ as to whether they reported that they had been given sufficient information (p=0.94); they had been able to make their own decision (p=0.21); staff had respected their decision (p=0.54); or that they were satisfied with their decision (p=0.94) and counseling received (p=0.82). The one provider-enhanced intervention assessed for its effect on satisfaction was the sequential cohort analysis that was previously described.^34,39 In this analysis, a higher percentage of participants in the intervention versus the pre-intervention cohort, having been read a short script on LARC effectiveness, reported that all their questions about birth control had been answered (p<0.001), and that they were satisfied with the discussion (p=0.03).

In combination with the original review,¹⁸ seven interventions were assessed for their effect on satisfaction/comfort with services and decision making (Figure 3, Appendix Table 2, available online). Findings remained mixed. Six of the seven interventions were provider-independent interventions, with three finding a significant positive effect,^33,48,60 two finding no effect,^26,51 and one finding a significant negative effect⁴⁰; only one provider-enhanced intervention was identified.^34,39

Two provider-independent interventions were identified in the review update that assessed attitudes toward contraception (KQ4), both with a positive effect. The first,⁴¹ evaluated through a pre-/post-test analysis of a comprehensive information brochure, as described earlier, resulted in an immediate post-test increase in participants’ scores for a positive attitude about OCs (p<0.001); although scores dropped after 3 months, they remained higher than at baseline (p=0.036). Moreover, the increase was positively associated with knowledge scores, both immediately post-test (r=0.284, p=0.001) and at the 3-month follow-up (r=0.206, p=0.022). The second intervention,³³ also described earlier, resulted in a significantly greater proportion of participants in the interactive Facebook group expressing a preference for implants in particular (p<0.01), or LARC in general (p<0.01). Moreover, knowledge was again an important factor. For both study arms combined, greater increases in knowledge (4%, 12%, 24%, and 36%) were associated with a greater increase in the RR of stating a preference for LARC (RRs: 1.06, 1.19, 1.43, and 1.71, respectively, 95% CIs in Appendix Table 2, available online).

In combination with the original review,¹⁸ six different interventions assessed the effect of educational tools on attitudes toward contraception (Figure 3, Appendix Table 2, available online). Except for one intervention using written materials alone,⁴⁶ each had a positive effect.^{33,41,46,58,61}

The review update identified four interventions, assessed through three RCTs^30,31,40 that addressed the selection of more versus less effective contraceptive methods and also provided education on the full range of contraceptive options (KQ5; Figure 3, Appendix Table 2, available online). Two of these interventions were provider-independent interactive tools described previously: one filled gaps in knowledge of LARC, while also providing information on the full range of contraceptive options,³¹ and the other presented methods in the order of effectiveness.⁴⁰ Neither resulted in differences in LARC uptake among intervention versus control participants (Gilliam et al.³¹: p=0.77, Sridhar and colleagues⁴⁰: p=0.75). The final study³⁰ found a positive effect of both a provider-independent and a provider-enhanced intervention. This study assessed a computer-based contraceptive assessment module in which intervention participants received either a generic list of methods (provider-independent), or a list of methods tailored to their responses, which they were instructed to share with their healthcare provider (provider-enhanced). Among participants in the control group who were not exposed to any tool, 65% selected an effective method (i.e., injectables, pills, patches, or rings), and 15% selected a highly effective method (i.e., IUD or implant); in comparison to these percentage for the control group, selection of effective and highly effective methods, respectively, was greater in both the generic (78% and 24%, p<0.001) and tailored information arm (75% and 22%, p<0.001), and both intervention arms had increased odds relative to controls of selecting an effective method (generic vs control: OR=1.74, 95% CI=1.35, 2.25; tailored vs control: OR=1.56, 95% CI=1.23, 1.98).

No studies from the original review¹⁸ evaluated the effect of educational interventions on selection of more effective methods.

The review update identified four interventions (two provider-independent^32,43 and two provider-enhanced^36,43), assessed through two RCTs^32,43 and one pre-/post-test analysis,³⁶ that asked whether a participant’s education level impacted knowledge gains (KQ6; Figure 3, Appendix Table 2, available online). One study assessed the interaction of patient education with use of text messages,³² one with written materials delivered by a provider using a flip chart,³⁶ and one with use of either a summary card or full-length brochure, delivered with or without interactive questions delivered by a provider.⁴³ With three of the interventions, knowledge gains were similar across education categories (text messages: p=0.49; summary card or full-length brochure without interactive questions: p=0.9; or with interactive questions: p=0.6).^32,43 Education interacted with the final intervention in which providers delivered information using a flip chart: Increases in knowledge were greater for participants with a high school education versus those with some college education or above (136% vs 51%, p=0.016).³⁶

No studies from the original review¹⁸ assessed the effect of educational attainment on knowledge gains.

This evidence summarizing the impact of educational interventions has several limitations, which should be considered when interpreting the evidence. Although several high-quality studies were identified for assessing the impact of educational interventions on knowledge, fewer were identified for the remaining primary outcomes targeted by contraceptive education. Seven studies assessed satisfaction/comfort with services and decision making, of which four were RCTs with low/moderate risk,^26,33,40,60 and three had high risk for bias.^34,39,48,51 Of the five studies assessing positive attitudes toward contraception, three were RCTs with low/moderate risk for bias,^33,46,58 and two had a high risk for bias or a less rigorous design.^41,61 Only three studies were identified that assessed selection of more effective methods, although all were RCTs with low/moderate risk for bias.^30,31,40 The authors were unable to identify any studies that addressed participation in the decision-making process. Additionally, as noted above, there was insufficient evidence to assess strategies for simplifying the presentation of information or the importance of input from a healthcare provider. Lastly, it is possible that additional articles meeting the inclusion criteria for this systematic review have been published since the updated search of the literature.