Contraception 88 (2013) 97–102

Original research article

Uptake of the levonorgestrel intrauterine system among recent postpartumwomen in Kenya: factors associated with decision-making☆

David Hubachera,⁎, Rose Masabab, Carolyne Kemunto Mandukub, Valentine VeenabaFHI 360, Durham, NC 27713, USA

bFHI 360, P.O. Box 38835-00623, Nairobi, Kenya

Received 14 December 2012; revised 1 March 2013; accepted 5 March 2013

Abstract

Background: The levonorgestrel intrauterine system (LNG IUS) may become more available in the public sector of resource-poor countries,but it is unclear what product features might be attractive to users and what factors will influence uptake.Study design: We recruited 671 women in Kenya who were seeking contraception at 6–12 weeks postpartum and gave them an opportunityto try the LNG IUS. We asked why they did or did not choose it, relative to the alternative options. χ2 tests of association were done toexamine participant characteristics and decision-making associated with choice.Results: Participants chose the following methods: LNG IUS (16%), injectable (36%), subdermal implant (30%), progestin-only pills (15%)and copper intrauterine device (IUD) (3%). Reasons for not choosing the LNG IUS included fear of pain/injury/discomfort (34%), modestyissues regarding insertion (33%) and fear of hormonal/health side effects (31%). Nearly a third of LNG IUS acceptors said they would havechosen a short-acting method if the LNG IUS were not available, and only 21% would have chosen the copper IUD.Conclusions: The LNG IUS could be an ideal method for increasing uptake of long-acting methods among recent postpartum women.Product attributes and comparisons to other contraceptive options are important factors in decision-making. Even among women comfortablewith intrauterine contraception, great distinctions and preferences are apparent. Addressing specific misconceptions and fears with betterinformation can help women make the best personal choices.© 2013 Elsevier Inc. All rights reserved.

Keywords: Levonorgestrel intrauterine system; Postpartum; Decision-making; Choice; Long-acting reversible contraception; Acceptability

1. Introduction

Short birth intervals increase the risk of maternalmortality [1] and adverse outcomes for the newborn [2];the United Nations Millennium Development Goals citeimproved access to contraception as a key action forrealizing improved health in resource-poor settings [3].New technologies may offer important advantages andincrease uptake.

☆ Funding for this project was provided to FHI 360 by the US Agencyfor International Development [GPO-A-00-08-00001-00, Program Researchfor Strengthening Services (PROGRESS)]. The views expressed in thispublication do not necessarily reflect those of FHI 360, the ICA Foundationor USAID.

⁎ Corresponding author. Tel.: +1 919 544 7040x11223, fax: +1 919544 7261.

E-mail address: dhubacher@fhi360.org (D. Hubacher).

0010-7824/$ – see front matter © 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.contraception.2013.03.001

Long-acting reversible contraception (LARC) consists ofintrauterine devices/systems and subdermal implants; col-lectively, these are the most effective temporary (spacing)family planning methods [4] and are safe to use in thepostpartum period [5]. Unlike injectables and oral contra-ceptives, LARC methods are easier to use because they donot require frequent periodic dosing and visits to healthcenters. In addition, the methods can provide between 3 and10+ years of contraceptive protection, depending on theproduct chosen. Thus, LARC can be a very attractive optionfor women who want to avoid early repeat pregnancies.

Historically, the copper intrauterine device (IUD) hasbeen the primary LARC method for public sector programsin sub-Saharan Africa. However, in the past two decades,prevalence of use has dropped or remained stagnant in manycountries. Currently, in the region, prevalence of IUD use isquite low (0.5%) compared to injectables (6.8%) and oralcontraceptives (4.3%) [6]. Just in the last decade, the

98 D. Hubacher et al. / Contraception 88 (2013) 97–102

subdermal implant has become increasingly popular (0.4%)and even outpaces the copper IUD in many countries(particularly those in Eastern Africa).

The levonorgestrel intrauterine system (LNG IUS) wasfirst approved in Finland in 1990. Worldwide productregistration has enabled millions of women to use it.However, the commercial product (known as Mirena®)that reaches sub-Saharan Africa (and developing countries inother regions of the world) is too expensive for the publicsector. The International Contraceptive Access (ICA)Foundation (a partnership between Bayer HealthcarePharmaceuticals and the Population Council) was formedin 2003 to donate and distribute free product known as “LNGIUS” to resource-poor settings. Since inception, the ICAFoundation has donated approximately 46,000 LNG IUSs to17 different countries [7]; this generosity has helped pave theway for new opportunities.

The LNG IUS is a proven technology that is arguablyoverdue for expansion into resource-poor settings. Themethod increases hemoglobin and serum ferritin levels [8];this could help alleviate chronic or acute anemia amongpostpartum (and other) women. In addition, because of near-universal postpartum/lactational amenorrhea, the postpartumperiod may be an ideal time for LNG IUS insertion; commonprogestin side effects may be either masked or alleviatedduring this time.

Though a new low-cost hormonal IUS is being developed[9–11], it will be years before it becomes widely available inthe public sector of developing countries. Thus, one interimoption is to offer the product to subgroups of the generalpopulation who are at highest need. We undertook thisproject (a) to provide postpartum women with anotherimportant method for preventing short birth intervals, (b) toassess how the LNG IUS might complement existing LARCoptions and (3) to characterize how the product is viewed.Together, the results can help prepare future efforts atintroducing the product on a wider scale.

2. Material and methods

We conducted a prospective cohort study in Nairobi,Kenya, and offered recent postpartum women the opportunityto try an LNG IUS. This paper addresses only the decision-making of the participants on the day they enrolled in thestudy; follow-up activities are ongoing.Womenwere recruitedat the Mathare North health center (operated by the NairobiCity Council and Kenyan Ministry of Health) from July 2011to May 2012. (This public sector facility is part of a networkthat serves marginalized populations in the metropolitan areaof Nairobi.) Participants enrolled voluntarily through aninformed consent process that was approved by the Protectionof Human Subjects Committee (of FHI 360) and the KenyaMedical Research Institute’s Ethical Review Committee.

Women were eligible for the study if they were 6–12weeks postpartum, 18–39 years of age, seeking a family

planning method, living in the clinic area and having accessto a phone. Women not meeting all the above criteria wereoffered standard family planning services. Approximately838 women were screened for eligibility. Of those who wereeligible (n=736), 671 enrolled voluntarily. As is customary atthis facility, women received free family planning methodsand services for all contraceptives including the LNG IUS.

The LNG IUS product donated by the ICA Foundation isessentially the same as the commercial product usedthroughout Europe and in the United States. It contains 52mg of levonorgestrel that is released from a reservoir on thevertical stem of the polyethylene device. The main differencebetween the commercial product and the LNG IUS is theinserter; the LNG IUS uses the linear plastic inserter,whereas the commercial product uses a curved inserter.

Women were told about all the contraceptive optionsavailable to them: condoms, lactational amenorrhea, proges-tin-only pills, the injectable depot medroxyprogesteroneacetate (DMPA), the levonorgestrel subdermal implant, thecopper IUD and the LNG IUS. The products were displayedand touchable. The study nurse explained differencesbetween the copper IUD (nonhormonal, lasts up to 10years, generally increases menstrual blood loss) and the LNGIUS (hormonal, lasts up to 5 years, sometimes causesspotting and usually decreases menstrual blood loss). Inaddition, women were told that the LNG IUS contains thesame type of medication that is found in many oralcontraceptives and in some subdermal implants, yet unlikethose hormonal methods, the LNG IUS medication worksdirectly in the uterus.

The nurse made clear that none of the options wasexperimental and all were approved for use in Kenya, even at6 weeks postpartum (regardless of breastfeeding status). Forany of the long-acting methods, women were told that theycould have the product removed at any time for any reason.Once participants had no further questions and they chose amethod, the nurse provided the selected method if nomedical contraindications were discovered. Then, the studynurse read the following text to participants (all of theaforementioned activities occurred on the same day):

During counseling, I described a new family planning methodcalled LNG IUS. You chose (insert her method here) that suitsyou best and that is fine. I do not want to change your mind. Ijust want to ask some questions about your decision.

The questions varied slightly, depending on the methodchosen (Table 1).

In response to the above questions, women were notgiven a list of possible responses, though most of theresponses were precoded on the form for ease of data entry.In addition to questions regarding participants’ contraceptivemethod decisions, we collected standard sociodemographicinformation as well as data on past/future pregnancies andprevious contraceptive method use. The data for this studywere double-entered in EpiData software. We calculated

Table 1Questions posed to different participants after they chose a contraceptive method

For participants who: Question

Chose the LNG IUS 1. What are the main reasons you chose the LNG IUS instead of the copper IUD?2. What are the main reasons you chose the LNG IUS instead of the subdermal implant?3. If the LNG IUS were not available, what method would you have chosen?

Did not choose the LNG IUS 4. What are the main features of the LNG IUS that make it unattractive or unacceptable to you?Chose the subdermal implant 5. What are the main reasons you chose an implant instead of the LNG IUS?Chose the copper IUD 6. What are the main reasons you chose the copper IUD instead of the LNG IUS?

Note: for all questions, multiple answers (reasons) were allowed.

99D. Hubacher et al. / Contraception 88 (2013) 97–102

percentage distributions and used Mantel–Haenszel and χ2

tests of association. All analyses were done using SAS.

3. Results

Of the 671 women enrolled, the following methods werechosen: progestin-only pills (14.7%), injectable (36.4%),subdermal implant (30.1%), copper IUD (2.5%) and LNGIUS (16.2%). Some background characteristics of partici-pants varied significantly according to the method that waschosen (Table 2). For example, pill users were lower paritycompared to other users, and injectable users were lesseducated compared to other users. Previous contraceptiveuse and desired length of method use were also associatedwith the chosen method. Of the participants who wanted totry the LNG IUS, 82% received it; in contrast, more than97% of participants who wanted one of the other methodsreceived their choice. [Reasons for not receiving the LNGIUS included medical contraindications (n=12) and lack ofspousal consent/personal readiness (n=8).] (Spousal consentis not required by law in Kenya.)

Among participants who chose the LNG IUS (n=109), thereasons for choosing it rather than the copper IUD or ratherthan the subdermal implant varied considerably (Table 3).Women cited the following reasons (in similar proportions)for choosing the LNG IUS instead of the copper IUD: onlyneed a 5-year product (47%), fewer side effects with theLNG IUS (44%) and less menstrual bleeding with the LNGIUS (43%). In contrast, when compared to the subdermalimplant, women were largely focused on just one difference:91% cited fewer side effects with the LNG IUS. Of the 109women who chose the LNG IUS, only 4 women hadprevious copper IUD experience and only 3 had previousimplant experience (data not shown); thus, the vast majoritywas not relying on personal experience to shape theirdecision-making.

The most cited reasons for not choosing the LNG IUS(Table 4) included fear of pain/injury/discomfort (34%),modesty issues regarding insertion (33%) and fear ofhormonal/health side effects (31%). The percentage citingspecific reasons varied somewhat depending on whichmethod was chosen by the participant. For example,women who chose subdermal implants cited modesty issuesto a far great extent than did those choosing pill. In addition,

fear of pain/injury/discomfort was a very important reason(40%) for not choosing the LNG IUS among those whochose the subdermal implant. In a subanalysis involvingparticipants who chose the pill, we reexamined the abovereasons by whether she had previous pill experience (n=47)versus not (n=52); the percent citing specific reasons did notvary significantly (data not shown). We did the samesubanalysis for those choosing DMPA and did not seesignificant variation by whether DMPA had been usedpreviously. Due to little previous implant use, the equivalentsubanalysis was not feasible for the participants who chosethe implant.

As noted previously, the number of women choosing thesubdermal implant (n=202) was about double the numberchoosing the LNG IUS (n=109). Subdermal implant userscited a wide range of reasons for choosing the implantinstead of the LNG IUS (Table 5). The most cited reasons forchoosing the implant included the following: less pain withinsertion/removal of implant (33%), fewer side effects(23%), insertion in arm is preferred to showing privateparts (22%) and the implant will not fall out or move aroundin the body (22%). Only 17 women chose the copper IUD;longer duration (10+ years) was the most cited reason forwanting it instead of the LNG IUS (data not shown).

About 30% of LNG IUS acceptors would have chosen ashort-acting method if the LNG IUS were not available(Table 6). Among the 70% that would have chosen a long-acting method, the subdermal implant was favored over thecopper IUD by 2:1. Of note, more LNG IUS acceptors wouldhave chosen the injectable (26%) compared to the copperIUD (21%).

4. Discussion

Sixteen percent of the postpartum participants wanted totry the LNG IUS. This decision was associated with varioussociodemographic characteristics and with specific beliefson how side effects and efficacy of the contraceptive optionsmight compare. Participants cited a variety of reasons forchoosing or not choosing the LNG IUS.

Women who chose the LNG IUS were presumablycomfortable with the notion of intrauterine contracep-tion; however, only 21% would have chosen the copperIUD as a substitute. This finding reveals the great

Table 2Distribution of background characteristics by contraceptive method chosena

Background characteristics Contraceptive method chosen

Pills (n=99) Injectable (n=244) Subdermal implant (n=202) LNG IUS (n=109) Total (n=654)

Agec

18–19 5.0 9.0 12.4 9.2 9.520–24 40.4 54.5 44.1 42.2 47.125–29 33.3 25.8 27.2 32.1 28.434+ 21.2 10.7 16.3 16.5 15.0

Number of childrenc

1 48.5 31.1 30.7 30.3 33.52 34.3 40.2 35.6 42.2 38.23 11.1 21.7 22.3 19.3 19.94+ 6.1 7.0 11.4 8.3 8.4

Currently married (%) 96.0 94.7 94.1 95.4 94.8Educationc

Less than primary 11.1 15.2 11.4 10.1 12.5Completed primary 51.5 63.1 54.5 51.4 56.7Completed secondary 24.2 18.8 29.7 27.5 24.5Higher 13.1 2.9 4.5 11.0 6.3

Currently breastfeeding (%) 100.0 98.8 97.5 100.0 98.8Previous use of a hormonal method or an IUDNo 35.3 41.4 32.2 31.2 35.9Yes 64.7 58.6 67.8 68.8 64.1

Previous use of chosen methodNo 52.5 52.9 95.5 100.0 73.8Yes 47.5 47.1 4.5 0.0 26.2

Method used before last pregnancyd

None 41.4 45.1 36.1 39.4 40.8Pills 30.3 10.2 19.3 13.8 16.7Injectable 11.1 36.9 27.7 33.0 29.5Other 17.2 7.8 16.8 13.8 13.0

Unintended last pregnancy (%) 33.3 43.4 49.5 43.1 43.7Last method used before unintended pregnancy

occurredb % and (n)N=33 N=106 N=100 N=47 N=286

No method 45.5 (15) 48.1 (51) 35.0 (35) 42.5 (20) 42.3 (121)Pills 24.2 (8) 14.2 (15) 27.0 (27) 17.0 (8) 20.3 (58)Injectable 9.1 (3) 28.3 (30) 22.0 (22) 23.4 (11) 23.1 (66)Condom 18.2 (6) 5.7 (6) 11.0 (11) 4.3 (2) 8.7 (25)Other 3.0 (1) 3.7 (4) 5.0 (5) 12.8 (6) 5.6 (16)

Ideal timing of next pregnancyd

Not sure 2.0 3.7 5.0 3.7 3.8Within 3 years 26.3 17.6 6.9 11.0 14.54–5 years 33.3 43.0 37.1 38.5 39.06+ years 16.2 16.4 17.8 18.3 17.1Never 22.2 19.3 33.2 28.4 25.5

Received chosen method (%)d 100.0 98.4 97.0 81.6 95.4a Excludes n=17 women who chose the copper IUD.b Among subset of women whose last pregnancy was unintended.c Mantel–Haenszel statistic for ordinal variable, p value b.05.d χ2 test of association p value b.05.

100 D. Hubacher et al. / Contraception 88 (2013) 97–102

distinctions women made when evaluating intrauterineoptions and the lower acceptability of the copper IUD;the preference for a 5-year versus 10-year duration ofaction was a key factor. Equally surprising, about 30%of women who chose the LNG IUS would have optedfor a less effective short-acting method as an alternativeif the product were not available. Thus, a sizeableproportion of LNG IUS acceptors were not committedto a long-acting method, but specifically wanted to trythe LNG IUS.

This research revealed many participant misconceptionsand fears that can be addressed in future introductoryactivities. Careful counseling and written (take-home)information for patients can be tailored to highlight themedical evidence that can counteract some myths. Forexample, many participants probably had exaggeratedconcerns over possible IUD expulsion and certainly devicemigration to other parts of the body. Rejecting the copperIUD because of the long duration (10+ years) is not a validreason for selecting the 5-year LNG IUS since all products

Table 3Why the LNG IUS was chosena instead of the copper IUD and instead of thesubdermal implant (n=109)

Reason the LNG IUS was chosen… Instead of the…

CopperIUD

Subdermalimplant

Fewer side effects with LNG IUS 44% 91%LNG IUS is more effective 24% 6%Less menstrual bleeding with LNG IUS 43% 4%The LNG IUS is expensive elsewhereand free now

3% 1%

Nobody will know I’m using the LNG IUS N/A 23%Only need a 5-year product 47% N/A

a Multiple reasons allowed; does not sum to 100%.

able 5hy the subdermal implant was chosen instead of the LNG IUS (n=202)

eason Percenta

ewer side effects with subdermal implant 24%ubdermal implant is more effective 8%ess pain with insertion/removal of subdermal implant 33%refer to expose arm for insertion rather than private parts 22%ubdermal implant won’t fall out or move 22%ubdermal implant won’t affect future fertility 1%ubdermal implant is well known and used by many women 11%ear of pain, injury or discomfort with LNG IUS 3%pouse will not or has not approved 1%oes not want device in uterus 11%

a Multiple reasons allowed; does not sum to 100%.

101D. Hubacher et al. / Contraception 88 (2013) 97–102

can be removed when desired. For some, it is possible thatfear of pain from an IUD insertion can be assuaged withinformation on scientifically measured levels of pain [12].Where possible, side effects need to be put in the propercomparative context. Some information exists on DMPAversus the LNG subdermal implant [13] and the copper IUDversus the LNG IUS [14,15]. Fear of contraceptive failurewas cited by many participants; the fact that all of the long-acting methods have similar, very low failure rates is animportant fact that would bear repeating in future introduc-tion activities. Some barriers to uptake, however, are moredifficult to solve or involve simple preferences: spousalissues, aversion to having a device in the uterus, modestyissues, etc.

Prior to our study, contraceptive uptake by postpartumwomen at the facility was minimal. For example, in a 4-month period preceding the study, 184 women presented for

Table 4Percent citing specific reasons why the LNG IUS product is unattractive orunacceptable to use, by contraceptive method chosena

Reason why LNG IUS wasnot chosen

Contraceptive method chosen

Pills(n=99)

Injectable(n=244)

Subdermalimplant(n=202)

Total(n=545)

Fear of pain/injury/discomfort 27% 33% 40% 34%Fear of hormonal side/healtheffects

36% 33% 25% 31%

Not familiar with product/notwell known/soundsexperimental

7% 13% 8% 10%

Method requires removalb 6% 15% – 12%Modesty issues regardinginsertionb

19% 34% 39% 33%

Does not want device in body 8% 7% 12% 9%Need to discuss with and/orget approval from spouseb

13% 16% 5% 11%

Fear that product will moveor fall out

10% 9% 12% 10%

Fear the method will failto prevent pregnancyb

4% 3% 10% 6%

a Multiple reasons allowed; does not sum to 100%.b χ2 test of association p value b.05.

TW

R

FSLPSSSFSD

well-baby services and 54 left with a contraceptive method(43 with DMPA, 7 with progestin-only pills, 4 with a copperIUD and 0 with an implant). In an average 4-month period inour study, we provided 230 women with contraceptives.Thus, our activity more than quadrupled contraceptiveprovision and increased (tremendously) the proportionreceiving a long-acting reversible method.

This research activity may represent the first public sectorfollow-up study in sub-Saharan Africa on client perceptionsof the LNG IUS; thus, this is an important step forward.Though many organizations have received donations andhave put the product into programs, research activities havenot been conducted. The focus on comparing the LNG IUSto existing contraceptive alternatives is another strength ofthis project. The decision-making information from thisreport will be supplemented with prospective follow-up data.Again, the focus will be on comparing the LNG IUSexperience to other method experiences based on overallsatisfaction, changes in bleeding patterns and productretention. The complete study will help inform the possiblerole of the LNG IUS for postpartum populations. Our efforthas two main weaknesses: the research was conducted in justone clinic, and the study size is relatively small. Thus, it isuncertain whether these results can be generalized to othersettings in sub-Saharan Africa.

Low cost LNG IUSs are also being developed in Belgium,China and India. It is conceivable that such products willenable millions of women to have affordable access to thisimportant technology. As more low-cost products becomeavailable, the world’s poorest women may have newcontraceptive choices to help prevent unintended pregnancy.

Table 6Contraceptive method the participant would have chosen if the LNG IUSwere not available (n=109)

Method Percent

Oral contraceptives 2.8Injectable 25.9Condoms 1.8Subdermal implant 48.2Copper IUD 21.3Total 100.0

102 D. Hubacher et al. / Contraception 88 (2013) 97–102

Acknowledgment

The authors would like to thank the ICA Foundation forproviding the LNG IUS for this project and the Nairobi CityCouncil andMinistry of Health staff at Mathare North HealthCenter for collaborating with us.

References

[1] Conde-Agudelo A, Belizan JM. Maternal morbidity and mortalityassociated with interpregnancy interval: cross sectional study. BMJ2000;321:1255–9.

[2] Cleland J, Conde-Agudelo A, Peterson H, Ross J, Tsui A. Contraceptionand health. Lancet 2012;380:149–56.

[3] The Millennium Development Goals report. New York: United Nations;2010. [cited 2011 March 3, 2011]; Available from: http://www.un.org/millenniumgoals/pdf/MDG%20Report%202010%20En%20r15%20-low%20res%2020100615%20-.pdf.

[4] WHO. Family planning handbook: a global handbook for providers.Baltimore and Geneva: Department of Reproductive Health andResearch Johns Hopkins Bloomberg School of Public Health Centerfor Communication Programs INFO Project; 2011. [cited 2012January 30]; Available from: http://www.globalhandbook.org/.

[5] World Health Organization medical eligibility criteria for contraceptiveuse. 3rd ed. Geneva: WHO; 2004 [cited 2010]; Available from: http://whqlibdoc.who.int/publications/2004/9241562668.pdf.

[6] World contraceptive use 2011. United Nations. [8 November 2012];Available from: http://www.un.org/esa/population/publications/contraceptive2011/wallchart_front.pdf.

[7] International Contraceptive Access (ICA) Foundation home page.International Contraceptive Access (ICA) Foundation. Available from:http://www.ica-foundation.org/.

[8] Kaunitz AM, Bissonnette F, Monteiro I, Lukkari-Lax E, Desanctis Y,Jensen J. Levonorgestrel-releasing intrauterine system for heavymenstrual bleeding improves hemoglobin and ferritin levels. Contra-ception 2012;86:452–7.

[9] Medicines360. Available from: http://www.medicines360.org/.[10] A study of a levonorgestrel-releasing intrauterine system for long-term,

reversible contraception. Clintrials.gov. Available from: http://www.clinicaltrials.gov/ct2/show/NCT00995150?term=medicines360&rank=1.

[11] Hormonal IUD. Uteron Pharma; Available from: http://www.odyssea-pharma.be/Hormonal-IUD.html?lang=en.

[12] Allen RH, Bartz D, Grimes DA, Hubacher D, O'Brien P. Interventionsfor pain with intrauterine device insertion. Cochrane Database SystRev 2009:CD007373 [Review].

[13] Hubacher D, Lopez L, Steiner MJ, Dorflinger L. Menstrual patternchanges from levonorgestrel subdermal implants and DMPA: systematicreview and evidence-based comparisons. Contraception 2009;80:113–8.

[14] Sivin I, el Mahgoub S, McCarthy T, Mishell Jr DR, Shoupe D, AlvarezF, et al. Long-term contraception with the levonorgestrel 20 mcg/day(LNg 20) and the copper T 380Ag intrauterine devices: a five-yearrandomized study. Contraception 1990;42:361–78.

[15] Andersson K, Odlind V, Rybo G. Levonorgestrel-releasing andcopper-releasing (Nova T) IUDs during five years of use: a randomizedcomparative trial. Contraception 1994;49:56–72.