A pilot quasi-experimental controlled trial of a community-based, multi-level family planning intervention for couples in rural Uganda: evidence of feasibility, acceptability, and effect on contraceptive uptake among those with an unmet need for family planning

Abstract

Objectives:

Effective interventions to reduce the unmet need for family planning in low-income settings are limited. This study aimed to establish the feasibility, acceptability, and preliminary effects of Family Health = Family Wealth (FH=FW), a multi-level intervention aimed to increase high-efficacy contraceptive uptake among couples wanting to delay pregnancy.

Study Design:

A pilot quasi-experimental controlled trial was conducted in rural Uganda with 70 couples wanting to delay pregnancy but not using contraceptives (n=140). Two matched clusters (communities) were randomly allocated to receive FH=FW or a comparator intervention via coin toss. FH=FW included health system strengthening elements and four facilitated group sessions. Interviewer-administered questionnaires were conducted at baseline and at ~7-months and ~10-months follow-up and process data gathered feasibility/acceptability outcomes.

Results:

Out of 121 households visited in the intervention community, 63 couples were screened, and 35 enrolled. In the comparator, 61 households were visited, 45 couples screened, and 35 enrolled. Intervention attendance was 96%, fidelity was 96%, and 100% of participants reported being satisfied with the intervention. From no use at baseline, there was 31% more high efficacy contraceptive uptake at 7-months and 40% more at 10-months in intervention vs. comparator couples (aOR = 1.68, 95% CI = 0.78-3.62, p = 0.19). A decline in fertility desires was observed in intervention vs. comparator participants from baseline (Wald χ² = 9.87, p=0.007; Cohen’s d: 7-months: 0.06; 10-months: 0.49).

Conclusions:

FH=FW is a feasible and acceptable intervention with strong promise in its effect on contraceptive uptake, to be established in a future trial.

1. Introduction

Reproductive autonomy (i.e., the power to control and decide about contraceptive use, pregnancy and childbearing) is a human right essential to the health and well-being of women [1]. However, in settings like Uganda, multi-level barriers contribute to a high unmet need for family planning (29.7% of married women who want to delay/stop future pregnancies are not using effective contraception) [2] and a high fertility rate (5.45 births per woman) [3].

Low individual knowledge [4-6], partner disapproval, poor communication and inequitable partner decision-making [6-8], community norms reinforcing gender inequity and large family size [6, 8, 9], and structural barriers (e.g., long clinic wait times, stock outs) are among the multi-level barriers to autonomous contraceptive use in Uganda [10-12]. Despite calls for multi-level interventions, few interventions take this approach, and there is a lack of interventions that have rigorous research designs (especially in Africa) that demonstrate effects beyond knowledge and attitudes [13-18].

The Family Health = Family Wealth (FH=FW) intervention is a multi-level, community-based intervention that aims to reduce the unmet need for family planning among Ugandan couples. Following stage 1b pilot study guidelines [19, 20], this pilot trial aimed to 1) evaluate the intervention’s feasibility and acceptability and 2) explore its effect on contraceptive uptake and fertility desires (desired number of children) among those wanting to delay pregnancy in the intervention community (relative to a water, sanitation, and hygiene [WASH] intervention) through ten-months follow-up.

2. Material and methods

From May 2021 to May 2022, we conducted a pilot quasi-experimental controlled trial to evaluate the FH=FW intervention vs. a comparator intervention. The study was approved by the University of Texas at San Antonio and the Makerere University School of Public Health institutional review boards and the Ugandan National Council for Science and Technology. A data safety monitoring board monitored the study. The study’s protocol has been published [21] and the trial was registered with Clinicaltrials.gov (NCT04262882) on February 10, 2020. The study took place in communities in a semi-rural district in central Uganda, located approximately two hours from Kampala. The population is 110,900 in an area of 270 mi². Family planning services are provided for free at decentralized governmental public health facilities (ranging across five levels: health centers I-IV, with variation in the availability of types of contraceptive methods by level). The Village Health Team, or the local cadre of community health workers, support community-based family planning efforts. Local private not-for-profits and private shops offer limited methods.

We selected communities based on 2020 population estimates of 249 villages across 5 sub-counties, first narrowed based on District Health Team recommendations. We then matched and selected two communities based on: population size (~2,000), distance to the health center V and having a health center III, no recent/ongoing community dialogues, similar religion and trade, and at least ten km from each other and not located along the same road to avoid contamination. We used coin toss to randomly allocate the two communities to intervention or comparator.

Research Assistants recruited participants through household mobilization, with support from the Village Health Team who helped identify households with couples of reproductive age, and gained permission for the research team to visit. Enrolled couples could also refer other couples to the study. The research assistant conducted eligibility screening, including pregnancy testing, and obtained written informed consent from both people in the couple. Village health workers, research assistants, and study participants were blinded to the study’s purpose (broadly framed as “family health”) and community allocation. Table 1 describes the inclusion criteria.

Table 1.

Summary of inclusion/exclusion criteria for couple enrollment in the Family Health=Family Wealth intervention pilot trial, Uganda 2021-22

Inclusion criteria
1	Married or considers themselves married
2	Couple living together most of the time
3	Living within the selected communities
4	Age 18 (or an emancipated minor) to 40 for women; age 18 (or an emancipated minor) to 50 for men
5	Luganda speaking
6	Woman not currently pregnant (self-reported and then confirmed with a pregnancy test for those meeting all other criteria)
7	At least one person in the couple reports wanting to avoid pregnancy for at least a year
8	Not using a method of contraception or using a lower-efficacy* or ineffective method of contraception based on success/failure rates of typical use (using condom less than 100% of the time, lactational amenorrhea, Fertility-Awareness Based methods (e.g., counting method), withdrawal method, spermicide, emergency contraception, sponge, and other traditional [herbal] methods)
9	Never used a non-reversible method (male or female sterilization)
10	Woman reporting having had sex in the past 3 months with spouse or planning to resume sex within the next 3 months with spouse if more than 1 month postpartum
Exclusion criteria
1	Pregnant as determined either by self-report or if found by pregnancy test
2	At least one person in the couple did not expect to be available for all sessions
3	Both the woman and man in the couple report wanting to get pregnant within the next year
4	The couple reported currently using an effective method of contraception (IUD, injection, oral pill, implant, vasectomy, tubal ligation, condoms 100% of the time)
5	The woman or man had known infertility (i.e., as told by a doctor)
6	Postpartum within 1 month of birth

Notes: Emancipated minor defined as those under 18 who are married, have children, or are pregnant. *Making an informed choice to use lower-efficacy methods is a person-centered family planning outcome. However, given considerable barriers to family planning use in the study area, these women were included in the intervention to ensure that this was an informed choice (rather than an outcome of inaccessibility to all methods and accurate family planning information).

2.1. Intervention conditions

The comparison village couples received a time/attention-matched WASH intervention [22], which was locally adapted in consultation with local health workers. All cohorts/groups received four total sessions, equal in structure and in duration to the intervention (see details in Table 2).

Table 2.

Overview of the Family Health = Family Wealth Intervention, organized by the three areas of “Family Health”: Physical Health, Relationship Health, and Economic Health, Uganda 2021-22

Sessions	Outlined content
Health System-Level
Pre-intervention health worker capacity building	Needs assessment conducted at public health facilities in intervention village to assess gaps in contraceptive knowledge and skills among health workers. Tailored family planning refresher training provided in partnership with the District Health Team to address training gaps.
Reduced wait time	Throughout the course of the study, intervention participants could go ahead in the queue for family planning with study ID card
Method distribution	Short-term methods provided at the end of sessions 3 and 4 (detailed below)
Couple Sessions
Session 1: Men’s Only Session Women’s Only Session ~90 minutes	Guided discussion to identify gender-specific definitions of “family wealth,” interpersonal and community barriers to family health and wealth, and redefine group norms on a “successful” family. Content tailored to the norms relevant to men and women’s separate groups. Program and family planning endorsed by a community leader
Session 2: Men’s Only Session ~2 hours	Relationship Health: Discussion on healthy relationships and family planning (partner violence, communication, decision-making, caregiver roles, gender norms); role modeling of gender equitable couples Economic Health: Business skill training co-facilitated with a local business expert (male expert) to increase interest in the program, improve couples’ shared decision-making, and highlight the importance of planning children to economic health
Session 2: Women’s Only Session ~2 hours	Physical Health: Contraceptive education co-facilitated with a midwife Economic Health: Business skill training co-facilitated with a local business expert, (female expert) to increase interest in the program, improve couples’ shared decision-making, and highlight the importance of planning children to economic health
Session 3: Couples’ Session ~2 hours	Physical Health: Contraceptive education co-facilitated with a Midwife; Midwife to provide family planning/linkages to care; create a “Family Action Plan” – setting family size and contraception goals Relationship Health: Communication skills building activities; create a Family Action Plan – setting relationship goals (take home assignment) Economic Health: Family budgeting
Session 4: Couples’ Session ~2 hours	Relationship Health: Communication skills building activity Revisit Family Action Plan goals as a couple Guided discussion to identify community barriers and solutions for family planning access/uptake Introduction to a “Community Action Plan” co-facilitated with community leader (e.g., Community Development Officer) Midwife to provide family planning/linkages to care Program and family planning endorsed by a community leader

Notes: Total of four sessions, two gender-segregated and two gender mixed. All sessions delivered by two trained intervention facilitators experienced in group facilitation and trained in the intervention, as well as in rights-based family planning and the importance of autonomous choice. Sessions take place approximately 1-2 weeks apart from one another. The WASH comparator was matched in timing and duration to the intervention.

Preliminary research from the study area [6, 7] and qualitative research conducted as the precursor to this trial informed FH=FW’s development. The intervention included two gender-segregated and two gender-mixed facilitated groups grounded in Campbell and Cornish’s social psychological theory of transformative communication [23] to guide communities to critically think about social norms underpinning a community problem, and reconstruct community norms together [24, 25]. This “community dialogue” approach was used to reshape norms reinforcing gender inequity and family size. Family planning’s benefits to “family health and wealth” (physical, relationship, economic health) was emphasized throughout. Dialogues were enhanced by integrating content informed by the social ecological approach [26, 27] to address multi-level barriers identified in the formative stage: health system elements (e.g., provider training, short-term methods offered during sessions); individual and couple-based education, skill building, and goal-setting activities; and community leader involvement; detailed in Table 2 and elsewhere [21]. A transport refund of 5,000 UGX (~1.50 USD) was provided for each session attended following local customs.

2.2. Data collection procedures and measures

Feasibility, or whether the intervention can be carried out and utilized [28], was obtained through study records and defined as the percent of planned sessions attended. Acceptability, or the degree to which participants were satisfied with the intervention [28], was collected through brief semi-structured interviews conducted after each intervention session. Fidelity, or the degree to which the intervention was implemented as prescribed in the protocol, was measured by the percent of planned intervention steps implemented, as determined by an analysis of randomly selected audio-recordings of 20% of sessions.

Aligned with FH=FW’s goal to help women make informed choices to use methods most effective for preventing unintended pregnancy, the main exploratory outcome is high-efficacy contraceptive uptake among those wanting to continue to delay pregnancy, collected through a computerized structured questionnaire following enrollment, and at approximately seven and ten-months follow-up in a private setting, or over the phone for follow-ups. Participants received 15,000 UGX (~4 USD) for each survey. Items adapted from the Uganda Demographic and Health Survey (DHS) [29] asked participants if they had used a method or done anything to delay/avoid pregnancy since the last interview, and if they were still using that or a different method. High efficacy methods based on typical use failure rates [30] and methods available/used locally (IUD, injection, implant, pills, tubal ligation, vasectomy) were coded as yes and lower efficacy methods (condom use, withdrawal, rhythm method, lactational amenorrhea method [LAM]) and no use were coded as no. LAM’s high efficacy reduces after 6-months [30]; based on the inclusion criteria, all women would have been greater than six months postpartum by follow-up. Self-reported use was confirmed, when possible, by participants’ family planning clinic cards, which detail receipt of services by the health facility. The variable was calculated as a between-dyads or couple variable (i.e., scores differ from dyad to dyad but not within a dyad) [31]. If there were couple discrepancies, the woman’s report was used.

The study’s secondary exploratory outcomes include unintended pregnancy incidence measured by pregnancy tests at baseline (part of study inclusion) and ten months. We classified pregnancy as “unintended” if participants stated that when they became pregnant, they wanted to wait until later or not get pregnant at all. We measured fertility desire at all time points using a DHS item, “How many more children do you want to have?” [29].

Other baseline items were assessed as potential covariates: age, household economic status (sum of DHS items on household characteristics, e.g., having electricity), polygamous marriage, tribe, religion, education, number of children, and lifetime contraceptive use.

2.3. Data analysis approach

We used descriptive statistics and frequencies to calculate feasibility/acceptability. We assessed baseline equivalence between the study arms on demographics using Generalized Estimating Equation (GEE) models in SPSS v. 28, which account for dependence from repeated, dyadic data. We tested intercorrelations between variables that differed (p < 0.10) between arms at baseline, or variables deemed theoretically relevant (i.e., age, number of children), to identify covariates. The time by intervention interaction on contraceptive uptake at follow-up was tested using GEE with a logit link distribution (couple the unit of analysis). We used linear GEE to test change in fertility desires (individual the unit of analysis), tested gender as a moderator, and calculated the Cohen’s d effect size for each follow-up [32]. We report unintended pregnancy descriptively (underpowered but included for feasibility). We conducted two sensitivity analyses of the time by intervention effect with different definitions of the outcome: any method use (regardless of efficacy) vs. no use, aligned with a person-centered approach, and any high efficacy use as defined by method failure rates based on perfect use (rather than typical use), making condom use high-efficacy [30].

3. Results

Figure 1 displays the CONSORT study diagram. Most households (94.5%) mobilized by the village health workers were visited by the study team, resulting in 182 couples pre-screened across both arms. Of those, 108 couples were screened for participation. After meeting recruitment targets (n=140 individuals, or 35 couples enrolled per village), 10 couples were withdrawn by investigators and replaced (7 intervention and 3 comparator couples) because at least one person in the couple failed to attend the first session. The participant-reported reasons for non-attendance are detailed in supplementary table 1. Including withdrawn couples (intention-to-treat [ITT]), study retention was 87.5% through ten months (82% intervention, 92% comparator). Retention among the final sample of 140 (excluding those replaced) was 99%.

Fig. 1.

CONSORT flow diagram of the Family Health = Family Wealth intervention pilot trial, Uganda, 2021-22”

Notes: *Couples where one or more person did not attend intervention session 1 were withdrawn after session

The average age of the sample was 29.8 years (SD = 7.59, range = 19-50). Most were of the Muganda tribe (87.9%, n = 123) and Muslim (58.6%, n= 82), which was higher in the control compared to intervention (p = 0.05). See Table 3 for sample characteristics.

Table 3.

Demographics at enrollment by study arm and by gender, with comparisons between study arms, Family Health=Family Wealth intervention pilot trial, Uganda 2021-22

	Full sample (N=140)	FH=FW Intervention			Comparator Intervention			Comparison between arms Test statistic
	M/SD or n (%)	Total (n=70)	Women (n=35)	Men (n=35)	Total (n=70)	Women (n=35)	Men (n=35)	OR/ b (SE) (95% CI)	p
Determinants of family planning
Couple-level variables (n=70)
Ever used high-efficacy contraceptives	53 (37.9%)	30 (42.9%)	20 (57.1%)	10 (28.6%)	23 (32.9%)	17 (48.6%)	6 (17.1%)	1.53 (0.71-3.29)	0.27
Socio-demographic variables
Couple-level variables (n=70)
Number of children	2.69 (2.04)	2.46 (2.02)	-	-	2.91 (2.06)	-	-	−0.46 (0.48)	0.34
Individual-level variables (n=140)
Age	29.89 (7.59)	29.01 (6.79)	26.51 (5.70)	31.51 (6.93)	30.77 (8.27)	27.83 (6.56)	33.71 (8.82)	−1.76 (1.51)	0.24
Wealth index sum	2.39 (1.16)	2.29 (0.95)	2.26 (1.01)	2.26 (1.01)	2.49 (1.34)	2.46 (1.42)	2.51 (1.27)	−0.20 (0.26)	0.45
In a polygamous marriage	22 (15.7%)	9 (12.9%)	4 (11.4%)	5 (14.3%)	13 (18.6%)	6 (17.1%)	7 (20.0%)	0.65 (0.19-2.24)	0.49
Tribe
All other	17 (12.1%)	9 (12.9%)	6 (17.1%)	3 (8.6%)	8 (11.4%)	5 (14.3%)	3 (8.6%)	1.14 (0.41-3.18)	0.80
Muganda (reference)	123 (87.9%)	61 (87.1%)	29 (82.9%)	32 (91.4%)	62 (88.6%)	30 (85.7%)	32 (91.4%)
Religion
Catholic, Protestant, and Other	58 (41.4%)	36 (51.4%)	17 (48.6%)	19 (54.3%)	22 (31.4%)	11 (31.4%)	11 (31.4%)	2.31 (1.00-5.32)	0.05
Muslim (ref)	82 (58.6%)	34 (48.6%)	18 (51.4%)	16 (45.7%)	48 (68.6%)	24 (68.6%)	24 (68.6%)
Education
Secondary or more	68 (48.6%)	38 (54.3%)	14 (40.0%)	24 (68.6%)	30 (42.9%)	19 (54.3%)	11 (31.4%)	1.58 (0.74-3.39)	0.24
Primary or less (ref)	72 (51.4%)	32 (45.7%)	21 (60.0%)	11 (31.4%)	40 (57.1%)	16 (45.7%)	24 (68.6%)

Abbreviations: M=mean, SD=standard deviation, b=unstandardized beta, SE=standard error, OR=odds ratio, CI=confidence interval

Notes: The statistical tests compare differences in demographics at enrollment between study arms (comparator=0 vs. FH=FW intervention=1). High-efficacy contraceptive uptake was based on success/failure rates for typical use, and included (based on what was available in the study area) IUD, implants, injectables, pills, and male condoms 100% of the time (those previously using tubal ligation/vasectomy were not eligible for the study).

In support of feasibility, nearly 99% of intervention session invitations extended were attended; using the ITT sample, there was 84% attendance. As for acceptability, all intervention participants reported being satisfied or very satisfied with the content of all sessions. Finally, 96% of all intervention steps were implemented (fidelity).

Table 4 reports intervention effects results, which control for religion and age. At seven months, 57.1% of couples in the intervention (n=20 couples) vs. 25.7% in the comparator (n=9 couples) were using a high efficacy contraceptive method (31.4% difference). At ten months, 57.1% (n=20 couples) of intervention vs. 17.1% of comparator couples (n=6 couples) were using a high efficacy method (40.0% difference) (see Figure 2) (aOR = 1.68, 95% CI = 0.78-3.62, p = 0.19). Between follow-ups, two intervention couples (10% of users) vs. three comparator couples (50% of users) discontinued use. Concordance in couples’ reports was high (seven-months: 100%, ten-months: 95%); 42% and 45% of self-reported use at seven and ten months were confirmed by clinic records. There was one reported unintended pregnancy in each arm at ten months. Finally, fertility desire was trending downward overtime for both arms (see Figure 3); the reduction was marginally statistically significant and greater (medium effect size) in intervention vs. comparator by ten months (7-months: b = −0.16, SE = 0.32, p = 0.60, Cohen’s d = 0.06; 10-months: b = −0.53, SE = 0.29, p = 0.07, Cohen’s d = 0.47) (no gender differences identified) and the arm by time interaction was statistically significant overall (Wald χ² = 9.87, p=0.007).

Table 4.

Family planning outcomes at baseline, 7-months, and 10-months follow-up by condition (unadjusted) and the arm by intervention effect (adjusted), Family Health=Family Wealth intervention pilot trial, Uganda 2021-22

Measured at the couple level	FH=FW Intervention (35 couples, n=69)			Comparator Intervention (35 couples, n=70)			Intervention Effect Arm*Time aOR (95% CI)	p
Primary outcome
High-efficacy contraceptive uptake among those wanting to delay pregnancy
10-months follow-up
Yes	20 (57.1%)			6 (17.1%)			1.67 (0.77-3.64)	0.20
No (ref)	15 (42.9%)			29 (82.9%)
7-months follow-up
Yes	20 (57.1%)			9 (25.7%)
No (ref)	15 (42.9%)			26 (74.3%)
Secondary outcomes
Unintended pregnancy
10-months follow-up
Yes	1 (2.9%)			1 (2.9%)
No (ref)	34 (97.1%)			34 (97.1%)
	FH=FW Intervention			Comparator WASH Intervention			Intervention Effect
Measured at individual level	Total (n=69)	Women (n=35)	Men (n=34)	Total (n=70)	Women (n=35)	Men (n=35)	Arm*Time b (SE)	Arm*Time Wald χ2
Fertility desire								9.87 p=0.007
10-months follow-up	1.83 (0.64)	1.74 (0.56)	1.91 (0.71)	2.36 (1.38)	2.14 (1.09)	2.57 (1.61)	−0.53 (0.29) p=0.07	Cohen’s D = 0.49
7-months follow-up	2.38 (1.24)	1.91 (0.74)	2.85 (1.46)	2.47 (1.40)	2.40 (1.44)	2.54 (1.38)	−0.16 (0.32) p=0.60	Cohen’s D = 0.06
Baseline	2.90 (1.49)	2.60 (1.38)	3.21 (1.55)	2.76 (2.59)	2.34 (1.59)	3.17 (3.28)

Abbreviations M=mean, SD=standard deviation, b=unstandardized beta, SE=standard error, aOR=adjusted odds ratio, CI=confidence interval

Notes: Models control for age and religion. Not using effective contraceptives at baseline and not being pregnant at baseline were part of the study’s inclusion criteria. The model for effective contraceptive uptake assessed change between 7-months and 10-months follow-up (baseline not included in the model). High-efficacy contraceptive uptake was based on success/failure rates for typical use, and included (based on what was available in the study area) IUD, implants, injectables, pills, tubal ligation, and vasectomy. Low-efficacy methods based on typical use reported in the study area included withdrawal, rhythm method, LAM, and male condom use. Fertility desire is one’s desired number of additional children.

Figure 2.

Uptake of high-efficacy contraceptive methods among those wanting to delay pregnancy through 10-months follow-up in couples receiving the Family Health = Family Wealth (FH=FW) intervention compared to couples receiving the water, sanitation, and hygiene (WASH) comparator intervention, Uganda, 2021-22

Notes: GEE model generated estimates controlling for age and religion. Error bars represent 95% confidence intervals. At baseline, no participants were using contraceptives (per study inclusion criteria); thus, this time point was not included in the model but is displayed in the figure for visualization. The couple was the unit of analysis (35 per treatment arm).

Figure 3.

Change in fertility desires (desired number of additional children) through 10-months follow-up in individuals receiving the Family Health = Family Wealth (FH=FW) intervention compared to individuals receiving the water, sanitation, and hygiene (WASH) comparator intervention, Uganda, 2021-22

Notes: GEE model generated estimates controlling for age and religion. Error bars represent 95% confidence intervals. Fertility desires are measured as the number of additional children desired. The individual was the unit of analysis 70 per treatment arm).

The types of methods used and information related to autonomous decision-making are presented in the supplemental table 2. Sensitivity analyses conducted with the outcome of any methods use (regardless of efficacy) and any high efficacy contraceptive uptake (based on perfect use) did not differ substantively from the main analysis (and are therefore not reported); there was only a small number of participants reporting lower-efficacy methods (see supplemental table 2).

4. Discussion

The findings of this pilot trial support the feasibility and acceptability of a community-based, multi-level family planning intervention for couples in rural Uganda. The preliminary effects on contraceptive uptake/continuation and fertility desires show promise that this intervention may significantly reduce the unmet need for family planning among Ugandan couples. Success was observed for both women and men, with equal retention, satisfaction, and reported change in outcomes by gender – a likely result of efforts made to tailor content to their specific interests.

The ability to carry out the intervention was supported by high participation and fidelity, and intervention satisfaction was high, supporting progression to a community efficacy trial, while informing modifications to improve implementation. Additional qualitative data to support these results will be published in a separate manuscript. Based on the withdrawal of participants due to non-attendance of session one, strengthening eligibility screening and making session one attendance the final stage of enrollment are possible modifications to increase the likelihood that those enrolled will attend.

Although it did not reach statistical significance, this trial shows promise that FH=FW can increase contraceptive uptake; there was 31.4% more uptake at seven months and 40% more uptake at ten months, which may suggest a clinically significant relationship, corroborated by the significant intervention effect on reduced fertility desires. However, as a pilot trial, these relationships need to be established in a larger trial. In addition, a future trial should establish the intervention’s effect on discontinuation; in this pilot, there was 40% less discontinuation between follow-ups in intervention compared to comparator couples, but the analyzable sample was small.

Cluster randomization was appropriate for this intervention to reduce the risk of contamination and to establish procedures for a future cluster randomized controlled trial. However, as a pilot, only two communities could be included. Although we matched communities, people in the intervention were more mobile/less available than the comparison, contributing to greater decline/ineligibility observed in the intervention (Figure 1). Further, we refined screening procedures in the intervention village before beginning enrollment in the comparison village, which may have also contributed to this imbalance. While analyses controlled for baseline differences, confounding is still a risk.

A post hoc analysis (supplementary table 2) identified differences between those withdrawn by investigators and those that remained in the trial (younger age, more in Muganda tribe, less lifetime contraceptive use). Further, more were withdrawn from the intervention arm (seven vs. three couples). Nonattendance may have been due to these differences or lower interest in the intervention. Those withdrawn reported wanting to remain in the study and mostly had work conflicts, but it is possible disinterest was the true reason for non-attendance among some. It is also possible that the participants we excluded due to unavailability stated that they were not available but were actually not interested (reducing the number of reported declines).

There was observed change in contraceptive use and fertility desires in both arms, suggesting factors beyond the intervention also influenced change. Possibilities include measurement reactivity, effects of the comparator intervention, or external factors, such as other family planning programs or the COVID-19 pandemic. Pathfinder International was implementing a multisectoral family planning program with health system strengthening (e.g., provider training) for all district health centers at the time of the study [33]. Participants reported receiving no other health-related dialogues beyond the study, but this or other ongoing programs may account for some observed change. Recall or social desirability bias could have influenced participants’ reporting of outcomes. Women’s reports were corroborated by their partner’s in nearly all cases. However, due to inconsistent use of family planning clinic cards at facilities, less than half of self-reported use could be confirmed with clinic records. Procedures for collecting clinic record data will need to be strengthened in a future efficacy trial.

FH=FW is a feasible and acceptable intervention with strong promise in its ability to increase contraceptive uptake among Ugandan couples wanting to delay pregnancy. This trial provides support for progressing this intervention to a community efficacy trial, which taking right-based approach, should also examine additional patient-centered outcomes [34, 35]. If found efficacious, this intervention would have important public health implications for Uganda, and potentially for other settings where a high unmet need for contraceptives is tied to gender norms, relationship inequity, community dynamics, and health-system barriers [36-38].

Implications.

The Family Health = Family Wealth intervention addresses multi-level family planning barriers through four group dialogues with couples paired with efforts to reduce health-system barriers. A quasi-experimental controlled trial provides preliminary support for its feasibility, acceptability, contraceptive uptake and fertility desire effects, and success in engaging both women and men.