Empirical lessons regarding contraception in a protracted refugee setting: A descriptive study from Maela camp on the Thai-Myanmar border 1996 – 2015

Conflict settings and refugee camps can be chaotic places, with large and rapid population movements, exacerbated public health problems, and ad hoc health services. Reproductive health care that includes family planning is of heightened importance in such settings, however, funding and resources tend to be constrained and geared towards acute health services such as trauma management and infectious disease containment. Here we report on the complexities and challenges of providing family planning in a post-emergency refugee setting, using the example of the largest refugee camp on the Thai-Myanmar border, in existence now for over 30 years. Data from 2009 demonstrates an upward trend in uptake of all contraceptives, especially long acting reversible contraception (LARC) and permanent methods (e.g. sterilization) over time. Increased uptake occurred during periods of time when there were boosts in funding or when barriers to access were alleviated. For example a surgeon fluent in local languages is correlated with increased uptake of tubal ligation in females. These data indicate that funding directed toward contraceptives in this refugee setting led to increases in contraceptives use. However, contraceptive uptake estimates depend on the baseline population which is difficult to measure in this setting. As far as we are aware, this is the longest reported review of family planning services for a refugee camp setting to date. The lessons learned from this setting may be valuable given the current global refugee crisis.


Introduction
Worldwide, an estimated 289,000 women died during pregnancy or childbirth in 2013. The overwhelming majority of these deaths were in disadvantaged populations, in rural and remote a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 camps in the late 1990s [18]. Maela camp was one of these consolidation camps and the population changed rapidly from an estimated 6,000 inhabitants in 1994 to 31,000 in 1998, growing to almost 50,000 in 2006, and is now approximately 38,000 as of December 2015, (based on Thai-Burma Border Consortium estimates) [19]. The camp is approximately 54 km from the next major town (Mae Sot, Tak Province, Thailand) and 20 km from the nearest Thai hospitals (Tha Song Yang Hospital to the north and Mae Ramat hospital to the south). The camp is ethnically diverse, but is primarily composed of the Karen minority group (mostly S'gaw and Pwo) as well as other minority ethnic groups from Myanmar. The main languages spoken in the camp are S'gaw Karen and Burmese followed by Pwo Karen.
Health and family planning services. Originally the main health provider in Maela camp was Médecins Sans Frontières (MSF) who withdrew in May 2005 handing over their camp facilities to Aide Medicale International (AMI). In 2011 AMI merged with Première Urgence (PU) creating PU-AMI who remained the major health provider in the camp until 2015. Commencing in 1986 Shoklo Malaria Research Unit (SMRU) initially conducted antenatal clinics (ANC) and birthing services alongside MSF with the original goal of preventing maternal death from malaria [20]. After discussions with MSF, SMRU became the sole provider of ANC and delivery services, including post-abortion care, in Maela camp in 1997.
Prior to 1991 women had access to a limited range of contraceptives through MSF medical outpatient clinics, including monophasic combined oral contraceptive pills (0.03 mg ethinyloestradiol (COCP)) and depomedroxyprogesterone acetate (150 mg, injected on a 12 weekly basis (Depo-injections)) but no family planning promotion or education and only brief counseling was offered in the outpatient department.
SMRU's family planning program began in May of 1996 in one of the original camps along the border (Shoklo) and in Maela refugee camp after consultation and permission from local Karen camp administrators. The amalgamation of the camps in the late 1990s into Maela required sudden expansion and increased demand on services and staff. For convenience, the SMRU family planning clinic in Maela was located adjacent to the ANC and birthing room and was staffed by trained male and female health workers. It provided a range of contraceptives as well as education and counseling to women and men, regardless of age and marital status. Contraceptive options included: natural methods, condoms, COCP, emergency contraception pills (ECPs), Depo-injections, Implant, Intra Uterine Device (Multiload-250 1 ); vasectomy in the camp and tubal ligation via mini-laparotomy in the Thai Public Hospital. The progestin only pill was not available in Thailand at the start of the program but it was made available from Family Planning Australia 1 (FPA) but uptake was poor and it was not continued. In 1996 Norplant 1 was the only implant available and in 2004 it was replaced by Implanon 1 . In 1996 we were unable to find any report of Norplant 1 use in refugees and a concerted effort was made to follow-up these women [21]. Misoprostol is not easily available in Thailand to humanitarian aid organizations. In early pregnancy SMRU uses misoprostol for treatment of fetal demise, annovulatory pregnancy and incomplete abortion; all confirmed by ultrasound. It is not used as an abortificant due to restrictive abortion laws in Thailand [22]. Rates of sexually transmitted infections (STIs) in Maela refugee camp, first examined by laboratory screening in 1997, are very low and there is no evidence they have increased over time in this population [23,24].
In April 2000, the Planned Parenthood Association of Thailand (PPAT) was funded and committed to provide comprehensive family planning services within Maela. In order to prevent service duplication, the SMRU core family planning staff and all SMRU clients and records were handed over to PPAT. PPAT reproductive health activities in the camp are significant and include family planning counseling to all post-partum women; outreach reproductive health education for men, women and adolescents in the camp and gynecological care (examination of breast and vaginal discharge, PAP smear, infertility and irregular menstruation counseling services). Reproductive health education from PPAT targeted men in order to promote male involvement in family planning, included women's training that focused on sexual and reproductive health and rights as well as peer education for adolescents. PPAT staff and volunteers also received specific training on ECPs.
SMRU continued to conduct vasectomy procedures in the camp until PPAT took over the provision of this service in April 2001. From 2005 through 2007 SMRU employed a surgeon who, with the assistance of a local midwife, established a female tubal ligation service on site, overcoming the need for referral to the Thai Public Hospital system [21]. Post-partum tubal ligation was the preferred method although interval sterilization was provided as needed. After the surgeon left SMRU (in 2007) provision of camp based sterilization service was only available when the attending obstetric doctor had that skill. PPAT were also able to refer to the Thai Public Hospital system.
Collaboration between the organizations has continued with PPAT intermittently conducting education at SMRU ANC and SMRU sharing birth lists of camp residents with PPAT who would conduct individualized follow-up visits. Also, PPAT collated monthly and yearly data for the whole camp while SMRU provided data on tubal ligation to PPAT and later on IUD.
Funding for contraceptives and family planning. Supply of contraceptives was initially fully supported by the Thai Department of Public Health (TDPH). The TDPH welcomed Karen staff from the refugee camp to receive training for Norplant insertion at a local hospital (Tha Song Yan Hospital). However, as a result of the 1997 Asian financial crisis [25] the TDPH was unable to continue to supply contraceptives. SMRU then sought funding elsewhere and FPA were able to provide an emergency budget for contraceptive supply for 36 months. PPAT has at times asked for donations from clients who can afford to pay to help cover the costs of contraceptives.

Data
To quantify the uptake of contraceptives and new family counseling visits, rates (e.g. use per 1,000 reproductive age women, per year) were calculated with the denominator being the baseline population. Pinpointing the current or yearly population can be difficult in refugee camps and resulting rates and figures require cautious interpretation [10]. Different agencies within refugee camps can provide overlapping services, so that isolating exact numbers without duplicate counting can be problematic. Maela camp is a closed population, meaning that residents are unable to legally travel within Thailand, however, the reality is much more complex [26] with refugees travelling out of the camp for economic purposes, in both legal and illegal capacities. Furthermore many villagers (non-camp residents) travel to the camp for health care and sustenance, as it is a source of food and basic medical services.
Two main organizations have kept population estimates for the camp: the Thai-Burma Border Consortium (TBBC) since 1998 and the United Nations High Commissioner for Refugees (UNHCR) since 2002. The TBBC provides services to refugees including some support for housing and food rations based on population counts [27]. TBBC numbers include camp dwellers who are not officially registered by UNHCR. UNHCR numbers started well after Maela camp was established and have systematically been lower than TBBC numbers but may be influenced by political sentiments; in this case they may have been influenced by real or perceived aggravation by Thai authorities regarding a growing, foreign, uninvited population, as with refugees elsewhere [28].
Since 1998, TBBC has reported monthly population estimates that include UNHCR estimates. We used population estimates from December of each year from both TBBC (1998 through 2015) and UNHCR (2002 and 2015).

Contraceptive uptake data
PPAT maintains monthly records of family planning consultations, including new and ongoing contraceptive consultations as well as house visits by PPAT for family planning counseling. "New" consultations are newly registered women to the service (SMRU or PPAT) and include women who started a method of contraception for the first time but not those who switched from a different method or had a gap in between using contraceptives.

Births and miscarriages per year
The vast majority of births in Maela camp occur at SMRU. All pregnant women providing a Maela address at ANC enrollment were included in the analysis. Pregnancy outcomes were noted regardless of place of pregnancy outcome and include live-and still-birth, miscarriage, or those lost to follow-up before pregnancy outcome was available (usually because of migration).

Analysis
For most years the population data are split by sex but not stratified by age. For years in which population by sex was not reported we estimated 49.79% of the population as female, using the average percentage of females for years in which the population by sex is reported. We then estimated the proportion of the female population who were of reproductive age (between the ages of 15 and 44) as 41.66% based on demographic surveillance systems in near-by Karen villages on the Thai side of the border [29].
We created smoothed population estimates by fitting a loess curve (a nonparametric locally weighted polynomial regression approach [30]) to the data points and extracting the smoothed values using R cran 3.2. These smoothed estimates were used as denominators for calculating the cumulative uptake of LARC and tubal ligation as well as the pregnancy and miscarriage rates. We used the smoothed population estimates from both TBBC and the UNHCR as a sensitivity analysis for our estimates of LARC and tubal ligation uptake, pregnancy rate and the miscarriage rate.
We calculated all new: contraceptive consultations; total pregnancies; and miscarriages for a given year per 1,000 reproductive age women. For LARC (IUD and implants) and tubal ligation we calculated the cumulative uptake per 1,000 reproductive age women. All calculations were repeated using both the TBBC and UNHR population estimates.

Ethics statement
For the extraction of data, ethical approval for retrospective analysis of pregnancy records was given by the Oxford Tropical Research Ethics Committee (OXTREC 28-09, amended 19 April 2012).

Population estimates
Population estimates for Maela camp are vastly different by reporting agency, with UNHCRbased estimates being consistently smaller than TBBC estimates (Fig 1)

Contraceptive uptake by type and year
Trends in contraceptive uptake were compared over time (Figs 2 and 3). Depo-injections, pills, and condoms have been the three major contraceptive choices in the camp (Fig 2). Depoinjections composed over 50% of all new contraceptives consultations until 2000 and remained the most popular method among new contraceptive clients in the camp until around 2008 when the pill began to make up approximately an equal proportion of new contraceptive type. Condom uptake from the clinic decreased. ECPs do not feature because there were no requests from clients for this method.
Cumulative uptake of long-acting and permanent methods IUD uptake has slowly increased throughout the time period (Fig 3). In 2002 cumulative IUD uptake was around 15 or 25 per 1,000 reproductive age women whereas in 2012 it was either around 125 or 75 per 1,000 reproductive age women (TBBC-and UNHCR-based population estimates respectively). Over the last decade, implant uptake appears to have at least doubled (TBBC-based population estimates) and potentially quadrupled (UNHCR-based population estimates). Female sterilizations (tubal ligation) followed the same general pattern as LARC until 2004 when they began to increase and surpassed LARC.
The cumulative uptake estimates here are sensitive to the underlying population estimates; however a spike in the uptake  reproductive age women having had the procedure), whereas by TBBC estimates it has increased slowly since 2007.

Pregnancy and miscarriage rates
As with the cumulative incidence of LARC uptake and tubal ligation, pregnancy and miscarriage rates are sensitive to the underlying denominator (Fig 3) Miscarriage rates showed small fluctuations over time but remained relatively stable for most of the time period (Fig 4). Using TBBC population estimates the miscarriage rate in 1998 was 21  increased to 30 per 1,000 by 2007 and began a decrease in 2011. There is an apparent decrease in recent years, with the most recent estimates (from 2015) of miscarriage rates indicating 10 or 11.5 per 1,000 women using TBBC or UNHCR estimates respectively.

Discussion
Our results indicate encouraging trends with regard to family planning in Maela refugee camp. LARC uptake appears to be increasing while pregnancies and miscarriages are decreasing (Figs 3 and 4). Pregnancy rates remain relatively high in the camp [31]. While pregnancy rates are high, the increased uptake of LARC and permanent methods coincide with decreases in maternal and neonatal mortality [32,33], though we are cautious in linking these trends because this analysis has not accounted for other mortality-related factors and the data do not allow us to draw causal relationships. Previous linkage between crude birth rates and mortality in refugee and internally displaced persons suggested reproductive health outcomes could be better in post-conflict phases than in the host country or in the country-of-origin [34].
The pattern of change reported here roughly follows the demographic transition model (with an initial decrease in mortality followed by a decrease in fertility) which has been observed in several regions of the world as populations undergo economic development [35,36]. Mortality data for refugees in relation to family planning services is frequently limited. Burundi refugees included in the 1988 census data in Tanzania had a 6 fold higher risk of child death compared to Tanzanian mothers who were at the highest risk of child bearing; the highest risk Tanzanian mothers having a child death was 6.4 fold higher than low risk Tanzanian mothers [37]. This emphasizes the need to provide family planning early on during refugee settings, when mortality is usually at its highest [38], when populations may be fluctuating most, and when the demographic and health provision situation may be most chaotic.
The stark increase in the uptake of female sterilization beginning in 2005 (Fig 3) coincides with employment of a surgeon by SMRU (2005-2007) who could provide tubal ligation immediately post-partum. Uptake was high because the doctor lived in the camp full time and could fluently speak the local camp languages (both S'gaw Karen and Burmese) allowing adequate explanation and strong doctor-patient relationships [39]. Local misconceptions and fears surrounding operations could be addressed directly rather than through an interpreter [40].
Miscarriages appear within the ranges reported from developed nations [41,42] and have remained relatively stable over time (Fig 4), with recent data suggesting declining rates. At the same time, ECPs uptake has been extremely low in the camp, despite education efforts geared toward both camp dwellers and family planning staff [43]. Karen culture is conservative with regard to intercourse and pregnancy [23,44] and low ECP uptake could be explained by sociocultural factors; by misconceptions about ECPs; or perhaps by women obtaining ECPs outside the camp [43]. The recommendation by Hobstetter et al for "increased outreach and education campaigns to raise community awareness and demand for ECPs" still stands but requires an appropriate level of funding, cooperation between agencies and addressing of misconceptions [43].
In conflict settings, decisions must be made about how and where to allocate limited funds [45] and reproductive health is not always prioritized. Changes in contraceptive uptake in Maela camp (Figs 2 and 3) are directly related to donor funding. For example, low literacy amongst the female population [46] required resources to be channeled toward training staff who speak local languages (S'gaw Karen, Pwo Karen and Burmese) in order to provide verbal or pictorial information and counseling to women. Qualified personnel (such as the previously mentioned surgeon) that could provide permanent contraception in a socio-culturally appropriate manner also required funds [39]. Free provision of contraceptives in the refugee camp probably contributed positively to uptake but data on the impact of patient donations was poor. Fluctuations in funding resulted in shifts in health services and the contraceptives that could be provided. This is well described in relation to other camp services by Cottrell [27] with regard to camp management and the provision of food. Likewise, the lower rate of implant uptake in 2000-2001 was related to the end of the FPA funding and in 2015 a decision was made to primarily provide IUD (since it permits at least 4 times more women to be reached compared to implants as the cost was 600 versus 2,200 Thai baht in the same year) because of the small family planning budget. Similar patterns are documented from other regions. For example, contraceptive uptake by Afghan refugee women living in Pakistan was observed to be directly associated with subsidized healthcare [47]. The call for agencies and governments to maintain funding for free family planning and other reproductive health services for refugees, internally displaced and marginalized populations is far from a new concept [48]. Recent massive population displacements, stemming from conflict in the Middle East [3] have refocused the spotlight on this important topic.
Contraceptive uptake in Maela camp is dominated by Depo-injections and the COCP (Fig  2). In the past, implant had higher uptake than did IUD but IUD uptake has consistently increased in the camp despite the slower uptake in different areas along the Thailand-Myanmar border [49] and Malaysia-Myanmar border [50]. In 2015 it was not possible to discern if higher IUD uptake was due to a required client donation for the implant or to uncertainty about the future as plans for voluntary repatriation of the camps move forward [51].
There are several limitations to this work. Foremost among these limitations are conflicting census estimates for Maela refugee camp. We have attempted to address (and highlight) this problem by using both UNHCR and TBBC population estimates as denominators. In relation to family planning programs in conflict settings Curry et al. [16] suggested the avoidance of costly population surveys and suggested that providers should focus on obtaining 'high quality routine monitoring data', using this to improve programs. The five sites studied by Curry et al all worked through the Ministry of Health unlike the situation in Maela camp [16]. In Maela camp census attempts and family planning uptake are obfuscated by the chaotic environment, complicated political situations, population fluctuations (frequently from new conflicts or skirmishes), limited funding and human resources. In places where the reproductive health needs are the greatest, the extent of those needs may be most difficult to accurately document and report, potentially leading to a cycle of public health and humanitarian neglect or misdirection of services [52].

Conclusion
The data presented here suggest that efforts to provide contraceptives in refugee settings are likely to yield benefits and that women and men, voluntarily seek such services when they are provided in an economically and socio-culturally appropriate manner. Unintended pregnancies represent unanticipated and largely avoidable social, economic, and health costs to women, couples, families and communities. Contraceptive services have significant and important benefits for maternal health and for communities [53] and humanitarian efforts should include contraceptives (not only ECPs). Funding agencies should give careful consideration with regard to the provision or withdrawal of family planning funding in humanitarian settings. Finally, data collection is a major problem in humanitarian settings and there is a great need for innovative approaches of collecting demographic and epidemiological indicators, given inherent complications as well as limited and fluctuating funding.
Supporting information S1