The Oxford Research Encyclopedia of Global Public Health has moved behind the paywall. For information on how to continue to view articles visit the how to subscribe page.
Dismiss
Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Global Public Health. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 29 January 2023

Experimental and Intervention Studies of Couples and Family Planning in Low- and Middle-Income Countries: A Systematic Reviewfree

Experimental and Intervention Studies of Couples and Family Planning in Low- and Middle-Income Countries: A Systematic Reviewfree

  • Stan BeckerStan BeckerBloomberg School of Public Health, Johns Hopkins University
  •  and Dana SarnakDana SarnakBloomberg School of Public Health, Johns Hopkins University

Summary

The vast majority of births in the world occur within marriages or stable partnerships. Yet family planning programs have largely ignored the male partner. One justification for this nearly exclusive focus on women has been that almost all of the modern contraceptive methods are female-oriented. In contrast, studies of fertility preferences within couples that included a later follow-up have shown that men’s fertility preferences are important for predicting subsequent births. Interspousal communication can be key to resolving differences in desired family size and for promoting open contraceptive use.

Experimental studies with couples on family planning education and/or counseling show higher contraceptive prevalence or continuation in the couples groups than in the women-only groups, though the differences are not always significant statistically. Other intervention studies have varying designs and mixed results. The purpose of this systematic review is to summarize the research findings on interventions with couples on reproductive health from experimental and pre–post observational studies. An important conclusion is that couples education and counseling are critical components for involving male partners. There is a need for systematic research on couples using a standardized intervention and fixed follow-up times and including analyses of cost-effectiveness.

Subjects

  • Global Health
  • Sexual & Reproductive Health

Introduction

Men and Couples in Family Planning

Human reproduction usually takes two persons.1 Though precise data are lacking, the vast majority of births in the world occur within married or in-union couple relationships. Thus, the couple would seem a natural focus for family planning research and programs. In fact, before 1960, prevention of pregnancy was very much a matter for couples—contraceptive methods primarily included withdrawal, the condom, periodic abstinence, and the diaphragm, which clearly involve male participation. However, since the introduction of the contraceptive pill in 1960 (and thereafter the intrauterine device and subsequently a multitude of other methods for women), medical science has focused on female methods of contraception. Why? There are three main reasons. First, there is one egg per month while there are millions of sperm per ejaculate. Second, because women give birth and are usually the ones to take children to health facilities for immunization and for health problems, they are in contact with the health system more than men, so there are more opportunities to talk to them about family planning. Third, because women are the ones who become pregnant, gestate, experience labor and delivery, breastfeed, and typically do most of the child care, they generally have a greater interest in preventing unwanted pregnancies. Patriarchy and related men’s sensitivities about their sexual pleasure and function vis-à-vis contraception (e.g., condoms and vasectomy) also play a role.

Whether to involve the husband or male partner is a perennial question that must be faced in family planning programs. From a clinical perspective, it is easier to treat a single individual, and because most contemporary contraceptive methods are for women, it is easiest to ignore the man. However, this strategy sometimes has negative consequences. Male involvement is crucial in the use of condoms for dual protection (Galsier & Crosignani, 2014). In certain contexts, if the woman uses contraception in secret and her partner discovers it, she may face retaliation such as emotional distress, marital dissolution, financial backlash and/or intimate partner violence (Bawah et al., 1999; Kibira et al., 2020).

While many women would like to involve their partners, some women may not! Since reproduction is not symmetric, patriarchy predominates in much of the world, and women are often the recipients of intimate partner violence, clinicians should only involve the male partner (e.g., for contraception, abortion, and sexually transmitted infections [STIs]) if the woman wants him included.

Still, given the important role men play—both as enablers and barriers to family planning—there has been a renewed interest in involving partners in current research and programs (U.S. Agency for International Development, 2018). Programs can target interventions toward men, husbands, or couples; the appropriateness of a particular approach varies by context.

Two-Sex Problem

Since contraceptive technology has focused on women, contraceptive programs have followed suit. Men have largely been absent from international family planning agendas despite them being consistently cited as both facilitators and barriers to uptake and continuation of family planning (e.g., Greene & Biddlecom, 2000). The United Nations acknowledged this “two-sex problem” in the 1994 International Conference on Population and Development held in Cairo; the final Programme of Action (Chapter VII) states: “. . . recognition of the basic right of couples and individuals to decide freely and responsibly on the number, spacing and timing of their children . . .” (United Nations, 1994/1995, p. 40).

Involving men and, more specifically, husbands has been theoretically and empirically shown to be effective in family planning research. Many studies highlight how men and women ideally see family planning as a joint decision and how it can even lead to marital stability (e.g., Harrington et al., 2019; Schwandt et al., 2021). Interventions targeting couples could seek to increase couple communication around family size goals and preferences, as well as encouraging safe and healthy negotiation and compromise where spouses differ. Interventions that encourage spousal trust may serve to counter the assumption made by some that contraception encourages spousal infidelity. Programs in Malawi targeting men found that raising awareness led to contraceptive adoption through the pathway of spousal communication within the couple (Shattuck et al., 2011). Male–male outreach and involvement of male champions for family planning has been shown as an effective way to change social norms and engender behavior change (Aransiola et al., 2014). Engaging men more generally in fertility and family planning issues, with the goal of changing gender norms in communities and societies, may help change more structural issues such as patriarchy.

However, involvement of men in family planning presents several measurement and programmatic challenges, especially in the context of unequal gendered power dynamics. Two individuals in a couple may have different goals regarding the number, spacing, and timing of their children, so some individuals cannot “decide freely” as the United Nations advocates, but in many cases must negotiate with their partner.

A related dilemma arises regarding unwanted fertility. If the woman says that she wants no more births, but her husband does, can one necessarily say the next pregnancy is unwanted? During a short course in Ouagadougou led by the first author, when this question was considered, it caused a major clash in the class, with the men saying of course it would not be unwanted as the wife “is carrying his baby,” and the women saying yes, it is her pregnancy and is clearly unwanted. The concept of unmet need for contraception has a related problem. Should unmet need only be measured for women? But if the woman has need but the husband does not, this can lead to covert contraceptive use, which, if discovered, could create major conflict in the union.

The importance of considering both spouses’ desires in fertility research is evident from studies which have asked each partner his or her fertility desires and then later determined if the couple had a birth or not. Figure 1 summarizes the results from eight such studies in low- and middle-income countries (LMICs), with data compiled by Cleland et al. (2020). In the figure, the reference value of 1.0 is for the proportion of couples who had a birth when both spouses said they wanted a birth. All of the 24 data points except one fall below 1.0. The ratio is lowest, as expected, when neither partner wants another birth. From the graph one can also tentatively conclude that the wife’s desires for another birth have slightly more influence than the husband’s desires, but clearly the husband’s desires are also important.

Figure 1. Ratio of the proportion having another birth for a given fertility preference group to the proportion having another birth for those couples with both wanting a birth, by group and country. Created by the authors using data from the review by Cleland et al. (2020). Countries included are the following: BD1 (Bangladesh: Razzaque, 1999); BD2 (Bangladesh: Razzaque, 1999); BD3 (Bangladesh: Gipson & Hindin, 2009); EG (Egypt: Casterline et al., 2001); MW (Malawi: Machiyama et al., 2015); MW1 (Malaysia: DaVanzo et al., 2003); MW2 (Malaysia: Tan & Tey, 1994); NG (Nigeria: Bankole, 1995).

Existing Research

Existing research on couples and sexual and reproductive health is difficult to summarize as there are many different research designs and outcomes. There are observational, cross-sectional studies of couples on matters related to fertility and family planning using survey data from husbands and wives separately (e.g., the Demographic and Health Surveys). There have also been longitudinal studies of couples tracking fertility preferences and subsequent fertility outcomes, as shown in figure 1. Less common, but potentially more meaningful, certainly for programmatic purposes, are couple studies that include an intervention. There are two subtypes of these—pre–post intervention studies and randomized controlled trials (RCTs), or experiments. The former includes the evaluation of a program, ideally with an intervention and comparison group, or possibly using a pre–post design of the same group. These studies need a longitudinal component to evaluate the effects of the intervention.

The second type includes experimental designs or RCTs. These include randomization to one or more intervention groups and a control arm. Again, there are several subtypes. In one, randomization is at the level of the clinic, community, or other geographic area. For some interventions, randomization at the community level makes the most sense—couples in some communities receive the intervention and couples in other communities do not. In the second, individual couples who consent to the study are randomized to one intervention or another; most studies in family planning clinics are of this genre.

Aim of This Study

Currently, there are systematic reviews that focus on the interventions of infertility treatments among couples (e.g., growth hormones [Sood et al., 2021], oocyte activation [Kamath et al., 2021] and metformin [Tso et al., 2020]), HIV prevention among couples (Hampanda et al., 2022), and couple studies on sexual and reproductive health more generally (Becker, 1996; Kraft et al., 2014; Nkwonta & Messias, 2019), but to date, there has been no systematic review of couple studies focusing on family planning interventions. Therefore, the purpose of this article is to systematically review the published literature on studies of family planning experiments and interventions evaluating uptake or continuation of contraception or use of abortion, which include both members of the couple and were conducted in LMICs. The article is limited to studies in these contexts as this is where the majority of fertility change and family planning research is being conducted. It is also restricted to studies of heterosexual couples. The focus is on intervention and experimental studies because results from such studies have minimal biases, resulting in more robust results than results from purely observational studies.

Methods

Search Methods

This systematic review proceeded as follows. An informationist from Welch Medical Library of Johns Hopkins University developed and conducted the literature searches, with team input. Searches were done of PubMed, CINAHL, and Scopus to identify studies published from the inception of the database to January 5, 2022. Search strategies were developed using a combination of controlled vocabulary and keywords to define the concepts of reproductive health, couples, and experimental and longitudinal studies. The PubMed search is presented in table 1.

Table 1. PubMed Search Strategy

1

(“Family Planning Services”[Mesh] OR “Contraception”[Mesh] OR “Fertility”[Mesh] OR “Abortion, Induced”[Mesh] OR “family planning”[tw] OR contracept*[tw] OR “birth control”[tw] OR fertility[tw] OR abortion[tw]) NOT (“Infertility”[Mesh] OR infertility[ti] OR infertile[ti])

2

“Spouses”[Mesh] OR couple[tw] OR couples[tw] OR couple’s[tw] OR spouse*[tw] OR spousal*[tw] OR partners[tw] OR partner’s[tw] OR (husband*[tw] AND (wife*[tw] OR wives*[tw]))

3

1 AND 2

4

“Cohort Studies”[Mesh:NoExp] OR “Follow-Up Studies”[Mesh] OR “Longitudinal Studies”[Mesh] OR “Prospective Studies”[Mesh] OR “cohort study”[tw] OR “cohort studies”[tw] OR “cohort analy*”[tw] OR “follow up study”[tw] OR “follow up studies”[tw] OR “prospective study”[tw] OR “prospective studies”[tw] OR longitudinal*[tw] OR panel[tw] OR experiment*[tw]

5

“Randomized Controlled Trials as Topic”[Mesh:NoExp] OR “Randomized Controlled Trial” [Publication Type:NoExp] OR “Random Allocation”[Mesh:NoExp] OR “Double-Blind Method”[Mesh:NoExp] OR “Single-Blind Method”[Mesh:NoExp] OR “Clinical Trial” [Publication Type:NoExp] OR “Clinical Trial, Phase I” [Publication Type:NoExp] OR “Clinical Trial, Phase II” [Publication Type] OR “Clinical Trial, Phase III” [Publication Type] OR “Clinical Trial, Phase IV” [Publication Type] OR “Controlled Clinical Trial” [Publication Type] OR “Randomized Controlled Trial” [Publication Type] OR “Multicenter Study” [Publication Type] OR “Clinical Trial” [Publication Type] OR “Clinical Trials as Topic”[Mesh] OR “clinical trial*”[tw] OR ((singl*[tw] OR doubl*[tw] OR treb*[tw] OR tripl*[tw]) AND (blind*[tw] OR mask*[tw])) OR “Placebos”[Mesh:NoExp] OR placebo*[tw] OR randomly allocated[tw] OR (allocated[tw] AND random*[tw])

6

4 OR 5

7

3 AND 6

Search results from each database were imported into the EndNote citation management system (EndNote Team, 2013), merged, and then imported into the Covidence systematic review software (Veritas Health Innovation, 2022) to remove duplicate records and facilitate screening and full-text review. Titles and abstracts were screened for inclusion; each author screened approximately half. Both authors simultaneously screened the studies eligible for full text review, and any conflicts were resolved through consultation.

Inclusion and Exclusion Criteria

Only studies from low- and middle-income countries’ (LMIC) contexts with an experimental or intervention design were included. A requirement for inclusion was having the intervention with both partners in a couple and/or the data were collected from this design. The main study outcome of interest was contraceptive use, that is, uptake or continuation, though some studies also reported on pregnancy or abortion. Studies had to be written in English, French, or Spanish. Studies were excluded if they were observational only or review, protocol, or pilot studies. Studies were also excluded if their focus was on method efficacy or on condom use to prevent sexually transmitted infections (STIs) and HIV. Finally, several studies were dropped that used data from randomized controlled trials (RCTs) but ignored the randomization and analyzed data for participants across the randomized groups.

Data Extraction

The following data were extracted from all experimental studies: country, region, and site; sample universe; design with sample sizes; description of intervention; follow-up time (longest if more than one); outcome or outcomes; findings; and quality score. The same data were extracted from intervention studies, except for the follow-up time. Both authors independently extracted the data on findings from the studies; conflicts were resolved through consultation. For the experimental studies, contraceptive use was the outcome of interest; some articles looked at all-method use and others only at modern method use. For comparative purposes, odds ratios (ORs) were compared. For some studies, these were provided by the authors; when they weren’t, the ORs were calculated from the two-by-two table of results by the authors. The log transform of these data was used and the respective standard errors were calculated for graphical display.

Quality Assessment

The studies included in the review met the above criteria; however, they varied widely in the strength of the design, description of the interventions, and interpretation of the results. Therefore, the quality of each study was assessed. A checklist similar to that of Sarkar et al. (2015) was used. For the experimental studies, the following criteria were assessed for each study to create a score out of 10: sample size justification; inclusion/exclusion criteria stated; intervention adequately described; dropout quantified; adjustment for lost-to-follow-up (LFU) done; data adequately described; data collectors independent of personnel doing the intervention; confidence intervals or p-values given; discussion of generalizability; and ethical clearance. For the intervention studies, a score out of 10 was also given, which included eight items from the same list, excluding the adjustment for LFU criterion and independence of data collectors, as well as two additional criteria: sample size >50 (per group); and adequate comparison group.

Analysis

Similar to other systematic reviews of family planning interventions (Cavallaro et al., 2019; Mwaikambo et al., 2011), the results of this search yielded studies with a wide range of target populations, study designs, and specific intervention approaches. Therefore, it was not appropriate to combine these data in formal meta-analyses. Instead, the studies have been summarized in four tables and a figure: two tables for the experimental studies and two for the intervention studies. The first table in each pair gives the location, design, sample size, randomization, intervention, outcome measure(s), and findings, and the second gives the data quality evaluation of each study. The figure gives the natural log of the odds ratio and its 95% confidence interval for intervention effects from the experimental studies. The articles predominately utilize p-values, with p < .05 taken as indicating a significant effect.

Results

Figure 2 presents the PRISMA Flow Chart. The search yielded 4,887 articles. After removing 748 duplicates, each author reviewed approximately half of the remaining 4,139 titles and abstracts to screen for eligibility. Of the 362 studies assessed for full text eligibility, each author independently assessed the papers against the inclusion criteria to determine whether these studies should be included in the review. Any conflicts were resolved through consultation. A group of 25 (16 experimental and nine intervention) articles was selected for data extraction. The authors knew of two other studies that the search had missed and they were included as well (Ashraf et al., 2014; Fisek & Sumbuloglu, 1978).

Figure 2. Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flow chart.

Experimental Studies

There were two subgroups of experimental studies. The largest group involved randomization of couple education and counseling about contraception; the second group had varying approaches to couples. Panel A of table 2 summarizes the 11 experimental studies with couple education and/or counseling. Seven were done in seven different African countries, three were done in Asia, and one in Latin America. Six had randomization at the community or cluster level, four had randomization in clinic populations, and one was individual randomization but in communities (see Column 3). Out of these 11 studies, seven found a significant positive effect of a couples education and counseling intervention on contraceptive use and four found no significant effect. Of the seven studies with the positive effects, the magnitudes of the differences in contraceptive prevalence at follow-up between the couples arm and either the women-only arm or the control arm (depending on the study) ranged from 3% in Malawi (Lemani et al., 2017) to 48% in Egypt (Soliman, 1999).2 It is also important to note the heterogeneity in follow-up times (see Column 5), which ranged from 2 months in the Guatemala study (Schuler et al., 2015) to 24 months in the Tanzania study (McCarthy, 2019).

Table 2. Summary of Published Articles (n = 16) of Experimental Studies of Couple Contraceptive Interventions in LMICs According to Type of Intervention

Region, country, and site (Study) (1)

Universe (2)

Design (sample sizes) (3)

Intervention (4)

Follow-Upa (month) (5)

Outcome (s)

(6)

Findings (7)

Quality score (8)

Panel A. Experiments with couple education and counseling on contraception

Africa

Burkina Faso; Bobo-Dioulasso five health centers in one health district (Daniele et al., 2018)

Women 15–45 in antenatal care; cohabiting, gestational age 20–36 weeks

Randomized to intervention (583) or control (561)

Three 1-hour educational sessions that took place during gestation and postpartum; one for male partners; two for the couple; three postnatal. Both 2 and 3 included FP

8

Use of effective contraception

At 8 months postpartum; 59.6% for intervention arm versus 53.1% for control (p < .05)

8

Egypt, maternity hospital at University of Mansoura (Soliman, 1999)

Women presenting at outpatient clinic who were in second trimester and some husbands

200 women; 100 husbands; randomized to two groups: (a) women and husbands, (b) control = routine care

Three counseling sessions (1 hour each) on FP methods

3

Contraceptive use at 3 months

91% in intervention versus 43% in control; p < .01

5

Ethiopia, Kotebe district of Addis Ababa (Terefe & Larson, 1993)

Married women of reproductive age

Two groups: (a) FP counseling with the couple (266), (b) counseling the wife alone (261)

Counseling session with a health assistant; content the same, maximum two visits

12

Modern contraceptive use at 2 and 12 months later

Use significantly higher in couple-counseling group versus wife-only group—33.0% versus 17% (p < .05)

5

Malawi, health center near Lilongwe (Lemani et al., 2017)

Women < 30 years; male partner; never used any modern contraception

30 health assistants randomized to receive (or not) the intervention (15 and 15)

(a) 2-days training in couples counseling, (b) control (no extra training)

6

Contraceptive continuation

No significant difference between women in two health assistant groups (but 94% and 97% in the respective groups, so very high)

8

Nigeria; Kano State; Antenatal clinic in teaching hospital (Abdulkadir et al., 2020)

Pregnant women age 15–45 in third trimester; husbands recruited or called

Intervention (75) or control (75) randomized individually

Two counseling sessions on contraception (one during antenatal before randomization and one 4 weeks later with spouses)

5

Acceptance of postpartum contraception

48.5% in intervention versus 31.0% in control p < .0001.

8

Rwanda, four districts (Doyle et al., 2018)

Men 21–35 years married or cohabiting and expectant or birth in last 5 years

Randomized individuals in 16 sectors (575 in intervention; 624 in control)

15 small group sessions (max. 45 hours) and their partners to eight sessions (max. 24 hours), met on a weekly basis; none for control

21

Contraceptive use

OR = 1.53; p < .004

9

Tanzania; Meatu district (McCarthy, 2019)

Married women 13–40 with husband in household

12 villages in three arms: (a) woman alone (4); (b) couples (4); (c) control (4)

Community-based distributor discussed birth spacing, contraceptive methods, etc. Intervention time period over 14 months but visits per household varied by village size (majority 4–6 visits)

24

Contraceptive use and pregnancy

For contraception, couples, treatment positive but only significant (p < .05) in one of three model specifications. Large negative effects of couples treatment on pregnancy (p < .01)

7

Asia

India, Maharashtra (Raj et al., 2016)

Husbands 18–30 years and their wives; infertile couples excluded

50 village clusters randomized to (a) three counseling sessions (469) or (b) control (612)

20–40 min sessions on maternal and child health and family planning, delivered in a 3-month timeframe

18

Modern contraceptive use

OR = 1.58; p = .023

9

Jordan, neighborhood of Aman (El-Khoury et al., 2016)

Married women, reproductive age, fecund, nonpregnant, not using contraception

Randomized to (a) woman-only arm (417), (b) couples arm (416), (c) control (414)

Counseling sessions—woman only counseling (home-based FP including vouchers) versus couple counseling (counselors trained on couples counseling and effective communication); vouchers; visits scheduled to accommodate husband’s schedules, over 6-month period

6

Modern contraceptive use

Uptake 8.5% higher in woman-only arm and 10.0% higher in couple arm relative to control (p < .01 for both)

7

Turkey, Istanbul; medical school hospital antenatal clinic (Turan et al., 2001)

Women after the fourth month of pregnancy

3 groups: (a) couples group (110), (b) women only (111), (c) control (112)

Four group educational sessions on (a) pregnancy, (b) birth, (c) infant care, (d) postpartum health and family planning, intervention was during the pregnancy

4

Postpartum contraceptive use

62% in couple group; 57% in women-only group; 47% in control group. Significantly higher in intervention groups (p < .05) but not between women-only and couples group

1

Latin America

Guatemala; 30 rural communities in western highlands (Schuler et al., 2015)

Married or civil union; woman 18–40 years

Two arms: (a) interactive workshops (166), (2) control arm (128)

Two workshops for women; two for men; two for couples, over the period of 1 month

2

Contraceptive use

Increase 10.9% in intervention group versus 4.0% in control but not significant

6

Panel B. Other experiments involving couples and family planning

Kenya, western (Harrington et al., 2019)

Women in ANC clinics in two public hospitals; 28+ weeks pregnant; access to mobile phone; HIV-negative

Randomized to 2-way SMS (130) or control (130)

Messages sent weekly until 6-month postpartum; partners too, over 6-month period

6

Highly effective contraceptive use

Relative risk increased for intervention group compared to control, 1.22 (1.01, 1.47)

6

Zambia, Lusaka (Mark et al., 2007)

HIV-concordant and discordant couples attending clinic; woman <38 years; fertile and nonpregnant

Couples randomized to three arms: intervention 2 (58); intervention 1 (111); and control (82)

Intervention 1—offered free contraception; intervention 2—intervention 1 + presentation about inheritance + wills; and control

3

Contraceptive use (at 3-month postintervention)

Contraception use much higher in two intervention arms (80% and 76% versus 33% in control) (p < .001); incidence of pregnancy not significantly different

7

Zambia, Lusaka (Ashraf et al., 2014)

Women 18–40 currently married, not pregnant, at one clinic in low-income area

Couples randomized to (a) couple arm (371) or (b) woman arm (378)

(a) With couple, husband given a voucher for free contraception, (b) woman given the voucher

24

Redeeming voucher

53% in individual arm and 43% in couples arm p < .05

Pakistan, Balochistan (Midhet & Becker, 2010)

All women in 32 villages

Eight village clusters women and husbands; eight women only and 16 control

Provision of information on safe motherhood, training of TBAs, husbands informed too in eight clusters

40

Modern contraceptive use

No significant differences

5

China, one district of Shanghai (Wang et al., 1998).

Nonsterilized married women in 21 factories and six schools

Nine work units in three arms: (a) wives only, (b) couples, (c) control (1,800 women total)

Educational intervention (e.g., five videotapes) in arms one and two

7

(a) Subsequent pregnancy or (b) abortion

No significant differences in total group, but for women not using IUDs, both pregnancy and abortion lower in couples than in wives only (p < .05)

7

Note: LMICs = low- and middle-income countries; OR = odds ratio; FP=family planning; IUD = intrauterine device.

a Some studies had two or three times of follow-up. Shown is the longest duration.

The type and intensity of the intervention varied across the experimental studies. In four studies, the intervention was education about contraception; other topics were also included. In four studies, both education and counseling were included, while three focused on counseling itself. The total number of education and counseling sessions varied from one or two in Ethiopia (Terefe & Larson, 1993), Nigeria (Abdulkadir et al., 2020), and Jordan (El-Khoury et al., 2016) to 15 in Rwanda (Doyle et al., 2018), with a median of three. The number of sessions with couples varied between one and eight, with a median of two (not shown).

Figure 3 gives the log odds ratios (and their 95% confidence intervals) for the effect of the couples intervention on subsequent contraceptive use in the 10 studies which gave the estimates or allowed for hand calculation of odds ratios by the authors, by time from intervention to follow-up.3 Another plot of the log odds ratio against the number of sessions with the couple did not show any significant trend, and in a regression of the log odds ratio with both the number of sessions and duration (in months) at follow-up, neither regressor was significant (not shown).

Figure 3. Logarithm of odds ratio (OR) (and 95% confidence interval) for contraceptive use in the couples experimental group versus women-only group (or control) by country and time (in months) of follow-up after the intervention. Triangle symbol indicates adjusted ORs; circle symbol indicates ORs. Blue color indicates modern contraceptive use; orange indicates all-method contraceptive use. Country studies are the following: GU (Guatemala: Schuler et al., 2015); EG (Egypt: Soliman, 1999); NG (Nigeria: Abdulkadir et al., 2020); TK (Turkey: Turan et al., 2001); MW (Malawi: Lemani et al., 2017); BF (Burkina Faso: Daniele et al., 2018); ET (Ethiopia: Terefe & Larson, 1993); IN (India: Raj et al., 2016); RW (Rwanda: Doyle et al., 2018); TZ (Tanzania: McCarthy, 2019). The one experimental study of the 11 in Panel A of table 1 that did not report an OR or the data for calculating one was the study in Jordan (El-Khoury et al., 2016). The study in Malawi adjusted variances for clustering and these could not be recovered from data in the article, so a confidence interval is not given.

What then characterizes or distinguishes the experiments with couple education and/or counseling that found a significant impact from those that did not? To start, only one of the three studies from Asia showed a positive effect (Raj et al., 2016) while five of the seven studies from Africa had significant positive effects (Abdulkadir et al., 2020; Daniele et al., 2018; Doyle et al., 2018; Soliman, 1999; Terefe & Larson, 1993). Further, the two studies with null results from Africa were somewhat unique. In the Malawi study (Lemani et al., 2017), couples were not randomized to a study arm; rather, randomization was one step removed; that is, at the level of health assistant training, those assigned to the couples intervention had a 2-day training on couples counseling. In the second study with null results from Tanzania, the results were not significant for the adjusted odds ratio in two of the model specifications, but the effect was significant in a third model (McCarthy, 2019). This study also showed a large and significant reduction in the proportion who reported a pregnancy at follow-up in the couples arm. Since women’s educational levels were higher in the Asian countries of this study compared to the African countries, it could be that male involvement had a greater effect where women’s average years of schooling were lower.

A second group of five experimental studies (Panel B of table 2) involved couples, yet differed from the first group in two ways. Four of the study’s interventions were not strictly administered to couples nor systematically compared with wives- or husbands-only interventions, and one study did not have contraceptive use as an outcome. The study in Kenya (Harrington et al., 2019) randomized pregnant women to receive (or not) SMS messages postpartum and there was an option to include male partners to receive SMS messages as well. There were no significant differences in contraceptive use between the groups at 6 months postpartum. The first study in Zambia (Mark et al., 2007) had three experimental arms, one of which was a control, but all groups targeted couples so there was no comparison to wives-only and/or husbands-only, as in the studies in Panel A. Contraceptive use reached very high levels in the two intervention arms and they were significantly different from that of the control arm. In the second Zambian study (Ashraf et al., 2014), the intervention was technically for couples but there was no education or counseling—in one arm, only the woman was given a voucher for free contraceptive services, while in the other, the voucher was given to the husband when both partners were present. A significantly higher proportion of vouchers were redeemed in the individual arm (p = .051). However, only 51% of women in that arm at follow-up responded that their overall health was good or excellent while the corresponding figure in the couple’s arm was 73% (p < .01); this may represent the psychological toll of covert use among some women in the individual arm. The study in Pakistan (Midhet & Becker, 2010) had female village volunteers who ran support groups for women in the villages and male volunteers simply distributed booklets and audio cassettes to husbands, so if couples were reached it was only indirectly. In the study from China (Wang et al., 1998), education was given on sexuality, reproduction, and contraception to husbands and wives separately (so there was no couple intervention per se), and pregnancy and abortion were the measured outcomes rather than contraceptive use. There was a significant positive effect on reducing pregnancies and abortions in the couples arm but only in the subgroup analysis of women not using intrauterine devices.4

Considering the quality of the experimental studies (table 3), out of a total score of 10, scores ranged from one in the Turkish study (Turan et al., 2001) to nine in the studies from India (Raj et al., 2016) and Rwanda (Doyle et al., 2018). Most studies clearly stated inclusion and/or exclusion criteria, described the intervention adequately, quantified the dropout rate, described the data adequately, presented p-values or other statistics, and discussed generalizability. There was more variation on sample size justification, adjustment for loss to follow up, and mention of ethical clearance.

Table 3. Quality Scores for 16 Experimental Studies of Couple Interventions for Contraceptive Use in LMICs

Study

Sample size justified

Inclusion/exclusion criteria stated

Intervention adequately described

Dropout quantified

Data collection independent of intervention

Adjustment for LFU done

Data adequately described

Confidence intervals or p-values given

Discussion of generalizability

Ethical clearance

Total Score

Panel A. Experiments with couple counseling on contraception

Daniele et al. (2018)

1

1

1

1

0

0

1

1

1

1

8

Soliman (1999)

0

0

1

1

0

1

0

1

1

0

5

Terefe & Larson (1993)

0

1

1

1

0

0

0

1

1

0

5

Lemani et al. (2017)

0

1

1

1

0

1

1

1

1

1

8

Abdulkadir et al. (2020)

1

1

1

1

0

0

1

1

1

1

8

Doyle et al. (2018)

1

1

1

1

0

1

1

1

1

1

9

McCarthy (2019)

0

1

1

1

1

1

1

1

1

0

8

Raj et al. (2016)

1

1

1

1

0

1

1

1

1

1

9

El-Khoury et al. (2016)

0

1

0

1

0

1

1

1

1

1

7

Schuler et al. (2015)

0

1

1

1

0

0

1

1

1

0

6

Turan et al. (2001)

0

0

0

1

1

0

0

0

0

0

2

Panel B. Other experiments involving couples and family planning

Harrington et al. (2019)

0

1

1

1

0

1

1

1

0

6

Mark et al. (2007)

0

1

1

0

0

1

1

1

1

6

Ashraf et al. (2014)

0

1

1

1

1

1

1

1

1

1

9

Midhet & Becker (2010)

0

1

1

1

1

1

1

0

0

6

Wang et al. (1998)

1

1

1

1

0

1

1

1

0

7

Note: LMICs = low- and middle-income countries; LFU = lost-to-follow-up.

One of the experimental studies of high quality was done in Rwanda (Doyle et al., 2018). This was the most elaborate of couple interventions as there were 15 sessions and included follow-up 9 and 21 months later. Attrition was quite low (14% at 21 months); the transportation stipend given of $2.50 each time may have been important in this regard. The authors conducted intent-to-treat analyses, with appropriate general estimating equation models to adjust for clustering. The adjusted odds ratio of contraceptive use in the intervention arm compared to the control was 1.53 (p < .004).

Intervention Studies

Table 4 summarizes the nine intervention studies. Four of them were conducted in African sites, while five took place in Asian geographies, with two each in Turkey and Bangladesh. The designs of these studies varied; while five of the nine included a comparison group or area, the other four simply had pre–post designs. Five were clinic-based and four were community-based. Five involved education and counseling with couples directly, and two had group education of either husbands and wives separately (Fisek & Sumbuloglu, 1978) or with couples (Jones et al., 2013).

Table 4. Summary of Nine Intervention Studies on Couples and Contraception in LMICs

Region, country and site (Study)

Universe

Design (sample sizes)

Intervention

Outcome(s)

Findings

Qual. score

Africa

Ethiopia, Oromia region; Jimma zone (Tilahun et al., 2015)

Married couples living together 6+ months, wife 15–49, not pregnant; monogamous

Three interventions (427) and three control districts (427); follow-up interviews at 12 months

Family planning education and promotion of husband–wife discussion on family planning; given to both men and women in household

Contraceptive use

47.6% use in intervention and 43.4% use in control (p > .05) but up from 41.9 to 47.6 baseline to follow-up in intervention.; in subgroup analysis there was increase in use among those not using at baseline (p = .014)

10

Kenya, Thika urban, peri-urban, and rural areas (Ngure et al., 2009)

Serodiscordant couples with 3+ episodes of intercourse in last 3 months; age 18+; CD4 count >250; not on antiretrovirals

Before and after and comparison (213) with other HIV clinical trial sites in Kenya

Training of staff on contraceptive methods; free contraceptives, involvement of male partners in contraceptive counseling session

Contraceptive use (noncondom)

Use rose among seropositive women from 31.5% to 64.7% OR = 4.0 (p < .05)

and 38.6% to 46.7% among seronegative women OR = 2.2 (p < .05); little change at other Kenyan sites (e.g., 46.7% versus 12.7%)

9

Rwanda; 4 health centers in and around Kigali (2 Catholic, 2 non-Catholic) (Mukamuyango, et al., 2020)

Women age 21–40 and men >20; fertile, not pregnant; not wanting conception in next 2 years, not using LARC

1,290 couples; no comparison group.

LARC offered at enrollment; motivational interviews 1 month later if not accepted before

Acceptance of LARC

74% accepted at enrollment; and another 6% accepted at 2nd visit; no comparison group for significance test

8

Zambia, Lusaka teaching hospital (Musaba et al., 1998)

Couples with one partner with symptomatic STD

Follow-up at 3, 6, and 12 months; no comparison group (99)

Couples received 60 male condoms and spermicides and 10 female condoms

Barrier method use at coitus

Less than 15% of acts were unprotected by a barrier method in 3, 6, and 12 months

6

Asia

Bangladesh, two districts (Khatun et al., 2011)

Married adolescent girls (and husbands)

279 women; 248 men

Training of service providers

Contraceptive prevalence

54% baseline to 57.3% comparison area; significance not given

8

Bangladesh, four family planning clinics, urban and rural (Amatya et al., 1994)

Women 18–40 years, previously pregnant, sexually active, etc.

Self-selection (by women) into husband counseling (408), i.e., bringing him to clinic or not (209)

Counseling of husband or not

Discontinuation at 36 months

32 per 100 in husband- counseled group; 42 per 100 in wife-only group (p = .07)

4

India, Chandigarh (Jones et al., 2013)

Couples “at risk” of STI, 18–59 years, monogamous, not pregnant, negative for HIV/STIs (30)

Three groups of 10 couples each with 3 weekly sessions (2 hour per session); all three groups had same intervention

(a) Safe sex, condoms, self-management, (2) sexual knowledge, sharing of experiences, role-playing

Condom use

42% in baseline to 100% in follow-up for men and 37% to 96% for women (p < .001)

5

Turkey, rural area north of Ankara; 31 villages (Fisek & Sumbuloglu, 1978)

Population in the Etimesgut health district

1,480 couples: three groups of villages: (a) couples; (b) wife only; (c) control; follow-up at 12 and 24 months

One-to-one education by auxiliary nurse-midwives; then group education

Contraceptive use at follow-up

Significant increase in contraceptive use in intervention area: 66% to 82% (p < .001) in wife + husband group; 65% to 75% in wife-only (p < .001) and 60% to 60% in control, but no significant difference between levels in the two intervention arms (p > .05)

6

Turkey, Gaziantep, three villages (Ozgür et al., 2000)

Women

Different intervention in each village; before/after as well

(a) Family planning education to women only (250 households), (b) family planning education to men only (325 households), (c) family planning education to both women and husbands (210 households)

Contraceptive use and effective contraceptive use

Unchanged, but rate of effective contraceptive use significantly increased in all arms, the most in group with education given to both sexes (p < .01 for before/after versus p < .05 for other groups)

4

Note: LMICs = low- and middle-income countries; OR = odds ratio; LARC= long-acting reversible contraception.

Of the five studies with a comparison group, the results were mixed. In the Ethiopian study (Tilahun et al., 2015), there was a significant increase in contraceptive use in the intervention compared to the control area among those not using contraceptives at baseline, though the overall difference in use at follow-up between areas was not significant. In the study of married adolescent girls in Bangladesh (Khatun et al., 2011), contraceptive prevalence increased from 54% to 57% in the intervention group but the increase in the comparison area was not given. In the Bangladesh study of Norplant (Amatya et al., 1994), the group with couples counseling had lower discontinuation at 36 months compared to the women-only group (p = .07). In the early Turkish study (Fisek & Sumbuloglu, 1978), contraceptive prevalence increased significantly in both the wife-only and couples groups, but the difference in those levels (75% versus 82%, respectively) was not found to be significant. In the second Turkish study (Ozgür et al., 2000), there were three intervention villages: one with family planning education to the wives only, one with the education to husbands only, and one with education given to both spouses. The increase in the proportion using effective methods was significant and highest in the village with education given to both. However, the authors did not test this against the increase in the villages with husband-only or wife-only education.

Regarding the quality of the studies (shown in table 5), out of a total score of 10, quality scores ranged from a low of four in the Turkish study by Ozgür et al. (2000) and the Bangladesh Norplant study (Amatya et al., 1994) to a high of 10 in the Ethiopia study (Tilahun et al., 2015). The intervention and data were adequately described across studies, and most included information on inclusion and/or exclusion criteria. However, quite a few failed to justify the sample size, include an adequate comparison group, report confidence intervals or p-values, or give information about ethical clearance.

Table 5. Quality Scores for Nine Intervention Studies of Couples and Contraceptive Use in LMICs

Study

Sample size

Intervention adequately described

Adequate comparison group

Inclusion/exclusion criteria stated

Drop-out quantified

Data adequately described

Confidence intervals or p-values given

Discussion of generalizability

Ethical clearance

Total score

Justified

>50 (per group)

Tilahun et al., 2015

1

1

1

1

1

1

1

1

1

1

10

Ngure et al., 2009

0

1

1

1

1

1

1

1

1

1

9

Mukamuyango et al., 2020

0

1

1

0

1

1

1

1

1

1

8

Fisek & Sumbuloglum (1978)

0

1

1

1

1

1

1

1

1

0

8

Musaba et al. (1998)

0

1

1

0

1

0

1

0

1

0

5

Amatya et al. (1994)

0

1

1

0

0

0

1

1

0

0

4

Khatun et al. (2011)

1

1

0

0

1

1

1

0

0

0

5

Jones et al. (2013)

0

0

1

0

1

1

1

0

1

1

6

Ozgür et al. (2000)

0

1

1

1

0

0

1

0

0

0

4

Note: LMICs = low- and middle-income countries.

Discussion

The purpose of this systematic review was to gather the evidence on family planning experiments and interventions that included both members of the couple and were conducted in low- and middle-income countries (LMICs). The best comparison group to evaluate this objective would be to compare results with those of the same or similar interventions that did not involve the couple (e.g., those with women only). In seven of the 11 experimental studies with couple education and/or counseling about contraception, the effect of including husbands and partners on contraceptive use when compared to a study arm with only women receiving the intervention was positive and significant. In the five studies that had different designs (e.g., only couple interventions, involving couples indirectly, or couples but with no education or counseling), significant effects were not documented. What seems clear is that the best approach to couples is to have education and counseling for them together.

A methodological note is needed on statistical significance and p-values. The studies and summaries given here relied on statistical tests, with p-values used to determine if the results were significant or not. Fortunately, most studies provided enough data so odds ratios and their 95% confidence intervals could be calculated, as shown in figure 3. It is well known that the threshold of p = .05 determining statistical significance is arbitrary.5 For example, in the Tanzania study (McCarthy, 2019), membership in the couples treatment group had a positive effect on use at follow-up in all three models but was only significant at p < .05 in one of them. Also, in the Zambia study, the negative effect of having both partners present when the voucher for services was given had a p-value of .049 in the model without controls and a p-value of .051 in the model with controls (Ashraf et al., 2014). It is important to point out that in the experimental studies without significant findings, the couple’s intervention group in all cases had greater contraceptive prevalence or continuation than the woman-only group (i.e., even when the p-value was >.05 [point estimates are included for the groups in the findings column of the tables]). One could argue that larger sample sizes would have found more significant differences but smaller differences thus detected would typically be less important programmatically.

A higher proportion of the experimental studies that showed significant positive effects of involving male partners came from sub-Saharan Africa contexts, and the median odds ratio for contraceptive use at follow-up among these five studies was 2.1. In that region, contraceptive use has been at low levels but has been increasing during the period covered by these studies.

The intervention studies also had designs that varied greatly. Unfortunately, one can conclude very little from the four that only had a before and after comparison because of the upward trend in contraceptive use, as noted. For the five studies with a comparison group (albeit nonrandomized), there were clearly significant results in two and, indeed, the group with the couple intervention had higher levels of contraceptive use at follow-up in all five studies.

Since the magnitude of biases in the observational studies could be considerable, the randomized trial is the preferred study design. It is noteworthy that randomization of interventions was not very burdensome in the studies examined here, so such experimental designs are recommended. However, the place of randomization in the design is critical. If women presenting at a clinic are randomized to have husbands invited to attend (experimental) or not (control), there would be immediate dropout from the experimental group of women whose husbands do not come for the intervention (or of wives who do not want their husbands involved). In such situations, intention-to-treat analyses would likely give very different results from a simple tabulation for those couples where the husbands did participate. For designs that involve visits to a couple in their home (e.g., El-Khoury et al., 2016), this is less of a problem but costs associated with home visits are higher.

How do these conclusions compare to other reviews of evidence around couple interventions and sexual and reproductive health or those done in different settings (e.g., high-income contexts)? First, though not included in this article due to scope issues, other reviews of evidence around couple interventions and prevention of HIV transmission have come to similar conclusions. A systematic review of the literature on couples HIV testing and counseling concluded that “couple-based interventions are more effective in promoting protective sex, HIV testing and Nevirapine uptake when directly compared to interventions delivered to individuals” (Crepaz et al., 2015, p. 1364). In the case of HIV, it is crucial for both partners to learn their HIV status, and in cases of serodiscordant couples, how to protect the HIV-negative partner and, for the HIV-positive partner, how to access treatment.

Second, this article has examined studies in LMICs and found support generally for couples interventions to increase contraceptive usage. How do these results compare with studies in developed countries? Actually, there is a paucity of comparable studies in those settings. One randomized trial enrolled couples in Los Angeles and Oklahoma City; couples were then randomized to receive either one informational session or three sessions addressing psychosocial factors and particularly couple communication (Kraft et al., 2007). The sessions had six to 12 couples. Contraceptive use 6 months later was not significantly different between couples in the two study arms. However, note that this did not test interventions for couples versus individuals. On the service provider side, there was a survey of family planning clinics in the United States, though only 45 of the 80 selected clinics agreed to participate. In data from clients in the clinics, it was found that 63% of the women were interested in couple counseling and one in five with a current partner had actually been accompanied to the clinic by him (Zolna et al., 2011). Given the dearth of studies in high-income countries, this may present an opportunity for innovation, where family planning programs in those countries can learn from the experiences reviewed here from LMIC countries.

Implications for Interventions

One fairly major hurdle encountered by programs that want to include male partners is their accessibility. Often men cannot accompany their wives to a clinic unless the clinic has evening or weekend hours as many are working during the daytime on weekdays. Furthermore, many men may feel uncomfortable in what is often perceived as a “women’s place.” Perhaps certain evenings can be set aside as “couple services.” Though most of the studies had retention rates of 80% or higher at follow-up, a major problem in a few of the experimental studies was the low level of couple participation in the intervention itself. For example, in the Egyptian study in the maternity hospital (Soliman, 1999), in only half of the couples were the husbands available for the intervention. In the Turkish study (Turan et al., 2001), only 26% of expectant fathers attended at least one of the education sessions. In the Guatemala study (Schuler et al., 2015), only 55% of enrollees in the intervention group both attended a session and had the follow-up survey. Thus, there is a clear advantage of interventions delivered in the home.

A second problem that programs targeting couples may face is around the psychological barriers of some male partners. Qualitative studies in Kenya, Uganda, Nigeria, and Ethiopia have shown that some men and women believe family planning to be a “woman’s matter” (Kabagenyi et al., 2014; Kibira et al., 2020; Withers et al., 2015). Studies have also highlighted how men may be reluctant to participate in family planning as it conflicts with traditional male gender roles. Specifically, it supposedly encourages infidelity or promiscuity of their wives and is controlled by their wives. Furthermore, they may be embarrassed or uncomfortable discussing sexual matters in front of or with their wives (Adanikin et al., 2019; Bawah et al., 1999; Kabagenyi et al., 2014; Withers et al., 2015).

A related question that arises is whether to have educational sessions on family planning and contraception with males separately from those for women or with the couple together, or a combination of these strategies. For example, in the Burkina Faso experimental study (Daniele et al., 2018), there was one educational session with male partners only and two sessions with the couples. In the Guatemala study (Schuler et al., 2015), there were two workshops with women, two with men, and two with couples. All but three of the experimental studies had direct couple counseling; the other three in India (Raj et al., 2016), Rwanda (Doyle et al., 2018), and Turkey (Turan et al., 2001) had couples education in groups. Qualitative studies that have explored this issue have also highlighted how family planning programs that target couples can help reduce misinformation and build trust within couples (Harrington et al., 2019; Withers et al., 2015). Note that sessions with women or men separately are easier to organize than sessions with couples, and several studies judged these separate sessions important to have in order to consider matters specific to each gender.

Also, in settings where women’s and men’s public spheres overlap very little, having couple counseling may not make sense. There is a related problem in patriarchal societies. In one qualitative study in Kenya, with researchers taping couple family planning counseling sessions, it was found that male partners dominated the discussions, with “77% of all client turns” (Kim et al., 2000). Thus, providers need to make sure a woman’s concerns are addressed. However, the same researcher noted: “Yet a woman may benefit from the presence of a male partner, even if it silences her, because he may raise issues that she would be reluctant to bring up herself and may elicit more information from the provider” (Kim et al., 2000, p. 46).

One of the initial hopes was that this review would help to inform cost-effectiveness of programs that involve couples versus those that do not. Family planning program managers in LMICs would clearly want to know the additional costs incurred to include male partners in their interventions. Unfortunately, only one of the articles reported added costs for the couple intervention; Fisek and Sumbuloglu (1978) estimated that the educational intervention with both spouses cost 49% more than the intervention with women alone. What would be desirable is a cost-effectiveness analysis; future studies could include this component. Indeed, the intensity of the interventions and therefore the costs varied considerably between studies, with only one couples session in the studies in Nigeria (Abdulkadir et al., 2020), India (Raj et al., 2016), and Jordan (El-Khoury et al., 2016) compared to eight sessions with couples in the Rwanda study (Doyle et al., 2018).

Recommendations for Research

There was wide variation in the interventions reviewed from one session in an antenatal clinic in Nigeria (Abdulkadir et al., 2020) to 15 sessions in the community in Rwanda (Doyle et al., 2018). There were also different structures of the interventions (i.e., sessions with wives alone, with husbands alone, or with couples). Is a session with the woman alone and then the partner alone and then together the best option?

This wide variation in designs leads the authors to recommend further research where WHO or some other international organization can convene an expert group to decide on the best design, including sample size, the intervention(s), and the follow-up period, as well as sponsor a multisite study. Such a study in five to 10 countries would have identical or nearly identical experimental designs with similar interventions, follow-up times, and analytic methods. This has been done previously (e.g., for a study of injuries in women attending emergency rooms, carried out in 11 countries [da Silva et al., 2015]). The research would also include a cost-effectiveness analysis, which would be very helpful for programs.

Recommendations for Programs

The mainly significant results from the couple studies reviewed suggest that family planning programs in LMICs consider ways to involve partners in their activities. Of course, cost-effectiveness analyses would be good to have before launching a couples component. In the meantime, incorporating family planning education with couples at the time of antenatal care may be a relatively inexpensive approach. In countries that have family planning fieldworkers who visit in the homes, including the husband or partner in education and counseling (when the woman wants it) would also be fairly easy, though it may involve more evening and weekend working hours for family planning workers.

Limitations

There are several limitations of this study. First, one of the search inclusion criteria was that the article be written in English, Spanish, or French; therefore, articles in other languages were missed, and potentially some in Spanish or French if the journal was not indexed in the search engines that were used. Studies from the gray literature were also excluded. Second, all studies reviewed used self-reports of contraceptive use, and such reports can be affected by social desirability bias, which has been shown to potentially be substantial in reporting of contraceptive use (Stuart & Grimes, 2009). For example, one might expect this bias to be higher in the intervention arm(s) because the respondents in those arms might feel a greater desire to please the interviewer, since they had both been given contraceptive information and/or counseling. Third, only published material from articles was utilized; by personally contacting authors of the respective articles and obtaining cost data, calculating cost-effectiveness estimates may have been possible, but this was beyond the scope of this article. Fourth, due to the varying designs, intervention approaches, and outcome measures, meta-analyses were not possible. However, odds ratios were available or could be successfully computed for most of the experimental studies.

Conclusion

Generally, studies have found that including male partners in contraceptive education and counseling leads to significantly higher acceptance and continuation rates. The 25 studies reviewed here had very different designs including, among other things, content and intensity of the intervention, whether there was a control or comparison group, and follow-up times. Therefore, a recommended way forward for a future endeavor would involve an international agency-sponsored multisite study with standardized design and the recording of costs so that cost-effectiveness analyses can be performed.

Acknowledgment

The authors thank Lori Rosman of Welch Medical Library for doing the searches of databases. We had multiple meetings to determine the best search strategies. She also helped us with Covidence software.

Appendix

Further Reading

References

Explanation of Codes for Tables 3 and 5

Table 3

Column

Description (of 1); otherwise = 0

1

If a sample size calculation was given or reported

2

Detailed description given of couples and individuals eligible and included in the study as well as exclusions

3

At least some details of the intervention are given; e.g., number and length of sessions or counseling, material covered in sessions, who gave the sessions

4

Number of dropouts given (e.g., refusals, moved)

5

As stated (of course, this is difficult in clinical studies)

6

Intention-to-treat analyses done or checking for differential dropout

7

Were questions from the questionnaire given and was there a description of data items collected and coding of these for the analyses?

8

Meaning is obvious

9

In the discussion section, generalizability of the findings was addressed

10

A specific institutional review board approval for the study is noted

Table 5

Column

Description (of 1); otherwise = 0

1

If a sample size calculation was given or reported

2

More than 50 couples or persons in each group (intervention and comparison)

3

At least some details of the intervention are given; e.g., number and length of sessions or counseling, material covered in sessions, who gave the sessions

4

Was there an appropriate comparison group studied?

5

Detailed description given of couples and individuals eligible and included in the study as well as exclusions

6

Number of dropouts given (e.g., refusals, moved)

7

Were questions from the questionnaire given and was there a description of data items collected and coding of these was performed for the analyses?

8

Meaning is obvious

9

In the discussion section, generalizability of the findings was addressed

10

A specific institutional review board approval for the study is noted

Notes

  • 1. Cases of IVF and surrogacy are exceptions.

  • 2. What can explain the extreme outlier of 48% in the Egyptian study (Soliman, 1999)? In that study at an antenatal clinic, though each method had higher prevalence in the experimental arm than in the control arm, the predominant method was the intrauterine device (IUD) and their insertions occurred soon after delivery. The very high prevalence at 3 months can be attributed in part to the high continuation rate of the IUD. Also, the control group did not receive education but “routine care,” which may have had minimal content about contraception.

  • 3. For eight studies, the outcome was contraceptive use, for one it was contraceptive continuation, and for one it was contraceptive acceptance. Also, several studies had multiple times of follow-up; in such cases the longest time was used.

  • 4. Post-facto subgroup analyses within RCTs are open to criticism (Rothwell, 2005; Schulz & Grimes, 2005).

  • 5. There are arguments for using likelihood methods that better represent the evidence in the observed data than do traditional hypothesis-testing methods (Royall, 2000), but none of the studies reviewed used them.