Despite decades of progress in improving the delivery and availability of family planning services, high levels of unmet need for family planning still exist in many countries.¹ This suggests that novel approaches are needed to extend access to family planning services to women and couples who desire to limit or space their childbearing but are not currently using contraceptives. The integration of family planning with other health services may be one such approach. Although integration may seem logical, the results of efforts to integrate child² or primary³ health care services with other services suggest that integration presents many logistic challenges and that caution is advisable. Shelton and Fuchs warn that the fragility of health systems in many countries can constrain effective integration of services.⁴ Therefore, an evidence base demonstrating the effectiveness of integration needs to be established before substantial investments are made in promoting integration as a means of fulfilling unmet need for family planning.

Numerous studies have examined the integration of family planning with programs for HIV/AIDS or other STIs.^5–9 Other studies have explored integration of health services (in some cases family planning services) with childhood immunization programs¹⁰ or intimate partner violence programs.¹¹ Another study reviewed the integration of primary health care services in general.³ To our knowledge, however, no comprehensive review has been conducted examining integration of family planning services with any type of health service. Therefore, we reviewed the literature to understand the current state of knowledge about the effectiveness of such integration.

METHODS

We searched the peer-reviewed literature to identify quantitative studies conducted anywhere in the world that evaluated the integration of family planning services with any other type of health service. We established, a priori, two inclusion criteria for the studies in this review. First, the study had to report a family planning–related behavioral or reproductive health outcome, such as contraceptive prevalence or service utilization. Second, the evaluation study had to have either a single-group pre- and posttest design or a two-group control or comparison design. These criteria were established to make it easier to assess the strength of evidence regarding the effectiveness of integration.

The search covered five databases that abstract peer-reviewed journals relevant to public health: PubMed, Embase, CABDirect,* PsycINFO and Social Sciences Citation Index. We selected these databases to capture a broad range of geographic and disciplinary areas. To search the databases, we used the combination of a family planning–related term ("family planning," "birth spacing," "birth control," "contraception," "pregnancy planning" or "pregnancy prevention") and an integration-related term ("integration," "bundling" or "twinning"). We limited the search to articles published in English from 1994 to mid-2009. At first, we conducted the search using a 10-year period (1999–2009), but because the number of eligible articles was small, we expanded the search period to 15 years.

Our search identified 581 unique abstracts, of which 80 were relevant to our topic. Two of us (ASK and LG) separately reviewed all relevant articles for inclusion, reaching consensus through discussion. Of the 80, many were commentaries or descriptive cross-sectional studies that did not meet the inclusion criteria. Ultimately, nine articles met both inclusion criteria and are included in the present analysis.^†

In reviewing the nine articles, we paid special attention to the integration intervention and its evaluation. We could not identify an agreed-upon set of terms describing the distinguishing features of integration efforts, so we named and defined several potential approaches to integration before conducting the review. In an iterative process, we then refined the terms and their definitions to reflect how integration was implemented in the various studies.

The intervention characteristics that we considered included the type of health service that was integrated with family planning, the location of the integration (clinic, community or both) and the integration approach (Web Appendix Table 1). We classified integration approaches into five categories: referrals, community partnerships, service coordination, cross-training and structural approaches. In referral-based interventions, staff members providing one type of service encourage clients to receive or obtain the other service(s) from a separate set of providers in a clinic setting. In community partnerships, staff who provide a clinic-based service refer clients to specially trained community members or outreach workers to obtain one or more other services outside of the clinic setting (and vice versa). Service coordination entails clients' receiving multiple services delivered by multiple providers at the same site during a single visit (i.e., the "one-stop shopping" model). Cross-training involves expanding the training of staff so that they are able to provide more than one type of health service to a client during the same visit. Finally, structural integration refers to changing the infrastructure and administration of the relevant health services in order to make them more seamless—for example, by creating a single medical record for mothers and infants or eliminating rules by which family planning services are offered only at certain times. More than one of these approaches to integration may be implemented simultaneously.

We used a tiered system adapted from the Guide to Community Preventive Services to rate the strength of each evaluation's design.¹² The Guide presents a well-established system developed by the Centers for Disease Control and Prevention and its partners for reviewing evidence-based public health information. Our adaptation classified the strength of a study's design as being at one of four levels: greatest, moderate, fair or least. Study designs categorized as having the greatest strength used "concurrent comparison groups and prospective measurement of exposure and outcome."¹² The moderate tier consisted of retrospective designs, including case-control comparison designs. We labeled studies that combined a single-group pre- and posttest design with a cross-sectional comparison group, or that used multiple pre- and posttests without a comparison group (i.e., time series), as fair. Finally, those that used "single pre- and postmeasurement and no concurrent comparison group" or a two-group cross-sectional design were considered the least suitable for contributing to an evidence base.¹² Studies with weaker designs, such as single-group cross-sectional, were not included in the review.

Finally, to examine the effect of the integration intervention on family planning–related outcomes, we assessed the magnitude and statistical significance of the reported changes in contraceptive prevalence or service utilization. If reported, the magnitude and statistical significance of changes in outcome variables for the other (nonfamily planning) health services were assessed as well. We also noted whether studies reported clients', providers' or community members' perspectives of the integrated services.

RESULTS

Of the nine articles included in the review, eight were published between 1999 and 2009. The nine studies suggest that the integration of family planning with other health services is taking place around the world, as Sub-Saharan Africa, Southeast Asia and South and Central America are all represented (Table 1). All of the studies were conducted in developing countries, although this was not a criterion for inclusion.

Location and Approach

The studies included both clinic- and community-based interventions, used a variety of approaches (Table 1) and were integrated with a wide range of other health services (Table 2, pages 192–193). Most frequently, family planning was integrated with some type of existing health service for mothers, children or both, such as the Expanded Programme on Immunizations. Other interventions integrated family planning with existing services for sexually active women, such as HIV/AIDS programs.

Two studies^13,14 went beyond the typical focus of family planning integration efforts by integrating family planning with broad health interventions. Lundgren and colleagues reported on an intervention in El Salvador that, in an attempt to expand family planning participation among men, integrated family planning education and referrals with an existing water and sanitation program that included home visits and community meetings.¹⁴ Amin and colleagues described a collaboration in Bangladesh between a nongovernmental organization and the local government to provide microcredit assistance and childhood immunizations along with family planning education and referrals and nonclinical family planning commodities.¹³

Although many interventions had an explicit community component designed to reach potential family planning clients who might otherwise be missed by a static, clinic-based program, the extent of this component ranged widely, from health workers making home visits to community mobilization. In addition, two studies combined clinic and community-based intervention strategies. In the first, Alvarado and colleagues tested an intervention in Chile that trained community health workers to make home visits to pregnant women; the intervention also involved structural changes in clinic procedures, so that mothers and infants shared a clinical record and had dual appointments during the postpartum year.¹⁵ In contrast, the Bangladesh study started as a door-to-door intervention but shifted to include static, clinic-based integration.¹³

The methods of delivering integrated services varied. The five studies that featured community partnerships followed the model of having specially trained community members or outreach workers provide information, education and nonclinical services, while providing referrals for clinical services; none mentioned having clinic staff members refer clients to specially trained community members or outreach workers for additional services.

In addition to the Chile study, two other interventions involved substantial structural changes in how services were organized and administered. In Niger, special family planning clinics were abolished, and services were instead offered at a broader range of times and locations.¹⁶ In Nigeria, an integrated management information system for family planning and prevention of HIV and other STIs was developed and implemented.¹⁷

Strength of Evaluation Study Designs

The study by Huntington and Aplogan was the only one that met the criteria for having the greatest strength of study design; it was also the earliest published (1994).¹⁸ Both of the studies that received a moderate rating used case-control comparison designs, although the quality of the comparison group differed between the studies: The intervention and comparison groups were more comparable in Ghana than in Chile, and the Ghana study used a random sampling strategy whereas the Chile study used convenience sampling.^15,19 Two of the three studies that received a fair rating^13,14 combined single-group pre- and posttest measurement with two-group cross-sectional designs by conducting baseline surveys only with potential participants but then including both participants and nonparticipants in the follow-up survey; the third used multiple pre- and posttest measurements (i.e., time-series design) but did not include a comparison group at any measurement point.¹⁶

Outcomes

•Family planning—related measures. Studies of interventions that focused on clinic-based components tended to report service utilization outcomes, while those with community-based components tended to report contraceptive use or prevalence data. Seven studies^13,16–21 reported improvements in a behavioral or reproductive health outcome related to family planning, although not all of these studies reported the significance of these improvements. The other two reported either mixed results in family planning–related outcomes¹⁴ or found no improvements in these outcomes.¹⁵ No study reported that integration had a negative impact on a family planning–related outcome.

The quasi-experimental study and one of the case-control studies reported statistically significant increases in the mean number of new¹⁸ and continuing¹⁹ family planning clients in intervention clinics, while the comparison clinics reported no such increases. In Pakistan, use of reversible, modern contraceptives was significantly greater among respondents in intervention areas than among those in comparison areas, even after adjustment for community, household and individual characteristics.²⁰ All three studies reported the statistical significance of their family planning–related outcomes.

Several other studies reported large percentage point increases in family planning–related outcomes. For example, self-reported contraceptive prevalence increased by 10–39 percentage points in India²¹ and by 25 percentage points in Bangladesh;¹³ the number of family planning visits in which condoms were distributed increased by approximately seven percentage points in Nigeria;¹⁷ and a composite index of contraceptive uptake increased from less than 1% to 23–29% in Niger (not shown)^.16 The statistical significance of these increases was not reported, however, making it difficult to assess their relevance.

Lundgren and colleagues reported mixed results in family planning–related outcomes.¹⁴ They found an increase in self-reported contraceptive use among male program participants, but not among female participants. Furthermore, at follow-up, contraceptive use did not differ between program participants and nonparticipants among either men or women. Moreover, because of the study design, the possibility cannot be ruled out that the changes in men's contraceptive use at follow-up were due to historical events in El Salvador instead of the intervention.

Alvarado and colleagues' case-control study of service integration in Chile¹⁵ did not find a positive effect on any family planning–related outcome. The contraceptive initiation rate in the intervention clinic was similar to that in the comparison clinics, despite the availability of a broader range of contraceptive methods at the intervention clinic. Furthermore, the intervention clinic initially offered contraceptives free of charge, while the comparison clinics charged a small fee. However, the intervention clinic did report significantly better outcomes than the comparison clinic on measures of exclusive breast-feeding and infant weight and length.

•Other outcomes. Some studies reported indicators of integration success and acceptability beyond those related to family planning (e.g., outcomes for other health services, client perspectives of the integrated services). All six of the studies that reported outcomes related to the other health services involved in the integration reported improvements in those outcomes.^{13,15,17–19,21} However, the significance of these findings was not always reported.

Similarly, all three of the studies that reported providers', clients' or community members' perspectives concerning the integration intervention found positive reactions (not shown).^14,15,18 In El Salvador, nearly all participants thought integrating family planning information into water and sanitation programming was beneficial.¹⁴ In Togo and Chile, most providers felt that integration enhanced service provision.^15,18 However, the studies that reported clients' and providers' reactions to the intervention provided only limited information; for example, they did not discuss providers' or participants' perceptions of the potential disadvantages and challenges of integration.

DISCUSSION

Evaluations of integration efforts have been conducted around the world, and these interventions have taken various forms. Most efforts focused on integrating family planning with other reproductive, sexual or child health services, but a few attempted to integrate family planning with services that extended beyond traditional health provision. Furthermore, no single model of integration exists. Although all of the interventions integrated, bundled or paired different health services with family planning, each did so in a unique way—the services that were integrated, their location and the approach to integration all varied.

This lack of consensus about the nature of integration complicates the dialogue about how best to do it. As discussions concerning the value of integration and its various forms continue, there is a need to refine definitions of integration and to clarify approaches. As an initial step toward a common taxonomy, we have presented terminology to classify the various locations and approaches. This taxonomy may need to be refined and expanded as our knowledge about integration increases; however, the use of common terminology should make it easier to develop an evidence base for integration and for providers, program implementation staff and researchers to share experiences.

Although it makes sense that integration should result in fewer missed opportunities for providing relevant services, and in potentially greater efficiency in service delivery, the data supporting these claims remain insufficient. Our study, like previous reviews of integration efforts,^3,7 found that the overall quality of the studies was often poor, as many had a single-group or cross-sectional design. Thus, although these studies contribute to the knowledge base about integration and can help generate hypotheses for future research, their contribution to establishing an evidence base on the effectiveness of integration in improving family planning outcomes is limited. Of note, however, is that two of the three studies with the strongest designs reported significant improvements in family planning–related outcomes.

While the general weakness of the study designs limits one's ability to establish causality, the consistency of the results suggests that integration has positive effects on family planning–related outcomes. Summarizing the overall magnitude of those effects is difficult, however, because of differences in approaches, study designs, outcome variables and reported information. This review highlights the need for well-designed, well-executed evaluation research to determine the effectiveness of integration and the best approaches to implementation.

Several factors may be contributing to the dearth of high-quality evaluation research on integration. First, in the wake of the 1994 International Conference on Population and Development, integration has often been assumed to be an effective strategy (although some have cautioned against this assumption⁴). In an environment where integration is thought to be an inherently positive model of service delivery, investing time and resources in high-quality evaluation may not be a priority. Second, donor commitment to family planning as a global public health issue waned during the years covered by this review,²² potentially further decreasing the resources available for evaluation research on family planning–related integration. Finally, integration interventions often include a community component, and designing and implementing evaluations that include adequate comparison or control groups in a community setting can be challenging. While this review included only a small number of studies, the three that had the strongest study designs were primarily clinic-based interventions; three of the four with the weakest designs were exclusively community-based interventions or had a significant community component.

The studies reviewed here were selected, in part, because they reported a quantitative outcome relevant to family planning programs. Other indicators of potential program success and acceptability are also important to understand. Although data were limited, most providers, clients and community members reacted favorably to integration efforts. Future research should include further assessment of clients', providers' and community members' perspectives on integrated services, such as their views concerning the conditions under which integration may or may not be feasible or acceptable (e.g., at fixed locations vs. outreach sessions), ways to avoid or offset potential costs of integration (such as long waiting times) and strategies for better integrating clinic- and community-based services.

Where reported, the other health services that were integrated with family planning showed improved or stable outcomes; none indicated a decline. Although none of the evidence reviewed here suggests that integration has a negative effect on other health services, having good data on all health services involved in integration interventions is vital. Ideally, integration would result in synergistic benefits for all services. For example, the study in Bangladesh integrated family planning with microcredit programs, which have been shown to reduce the economic and social vulnerability of women; such vulnerability often leads to risky sexual behaviors.²³ It is likely that women who have become empowered through microcredit programs are more likely than other women to use family planning services, and that improved access to these services would further reinforce the empowerment of microcredit program participants. Yet to confirm this, good quality data on utilization of all services involved are essential.

Furthermore, none of the studies we reviewed addressed the issue of cost-effectiveness. Integration is generally assumed to reduce costs, especially if it involves service coordination or cross-training. Yet, the lack of data on the efficiency and costs of integration is glaring. Before interventions are adopted on a broad scale, we need to know whether integration makes economic sense and whether certain integration strategies are more cost-effective than others.

The results of this review are subject to two main limitations. First, the search was limited to articles published in English. Additional evaluations of family planning integration efforts may have been published in other languages. Second, because this review focused on peer-reviewed literature, the potential for publication bias exists. Studies with significant findings are more likely than those with negative results to be submitted for peer review and, as a result, to be published.²⁴ Therefore, the pattern of results reported here may overestimate the true effect of integration interventions.

In conclusion, current evidence suggests that integration of family planning with other health services may be beneficial. However, before integration is embraced as a broad-scale strategy to address the high levels of unmet need for family planning around the world, well-designed and properly executed evaluation research is needed to establish an evidence base for integration. Future research should investigate providers', clients' and community members' perspectives on the integrated services, report health outcomes for all services being integrated and evaluate the cost-effectiveness of integration. Understanding the impact of various integration strategies on expanding the reach of services, on the efficiency and costs of service delivery and on the overall impact on both family planning and other health outcomes would be extremely useful. The potential for benefit is clear, but the evidence base for whether and how to most effectively implement integration to obtain those benefits remains weak.