Severity of Abortion Complications in Zambia: Does economic status matter? Ann Moore 1, Mardieh Dennis 2*, Giulia Greco 2, Akinrinola Bankole 1

Transcription

1 Severity of Abortion Complications in Zambia: Does economic status matter? Ann Moore 1, Mardieh Dennis 2*, Giulia Greco 2, Akinrinola Bankole 1 1 Guttmacher Institute, 125 Maiden Lane, 7 th Floor, New York, NY London School of Hygiene and Tropical Medicine, Tavistock Place, London, WC1H 9SH, UK * Formerly Population Council Zambia during the time the fieldwork for the study was conducted Background Although abortion can be legally provided in Zambia on the grounds of health, economic distress, and rape (Zambia Legal Information Institute, published 1972), it is rarely performed by trained providers under the allowed conditions. Lack of awareness among providers and women alike about the situations under which abortion can be legally provided remains one of the primary barriers to the provision of safe, legal abortion. Many women turn to illegal, unsafe abortions. Stigma and economic concerns about accessing abortion play a part in influencing the type of abortion women choose, thus often determining the risks that they will face as a consequence of undergoing an induced abortion. Induced abortion within the limit of the law is supposed to be offered free of charge at public health facilities. Marie Stopes Zambia (MS Z), a non governmental organization, has clinics in Zambia that charge women for induced abortions on a sliding scale. In addition, MS Z has a network of private franchised clinics, otherwise known as Blue Star clinics, which have undergone a process by which their work has been evaluated, the provider may have undergone additional training, and they have been provided necessary equipment so as to be able to provide safe abortion services according to MS Z stipulations. All health care facilities in Zambia should be able to treat postabortion care, with more serious complications referred to higher level facilities. Medical abortion (MA) using mifepristone and misoprostol was introduced into Zambia in 25 public health facilities in Misoprostol, as a drug for preventing postpartum hemorrhage, is more accessible than the drug combination regime of both mifepristone and misoprostol. While expansion of MA services has been slow (Hendrickson et al. forthcoming), is becoming increasingly available in Zambia. While pharmacists are legally allowed to only sell it with a prescription, in a study of 76 pharmacies in 2011, 72 percent of pharmacists offered to sell mystery clients misoprostol; only 22 percent asked the clients to present a prescription and only eight percent provided the correct dosage information (Hendrickson et al. forthcoming). Therefore, misoprostol is in use both at public and private health facilities as well as through informal channels but it is not universally available nor it is not always being used correctly. This study documents decision making regarding abortion and postabortion care seeking behavior including costs of seeking abortion care among women in two districts in Zambia: Lusaka and Kafue. This is a longitudinal study meant to assess the household costs (health, economic and social) of abortion in Zambia. This work is part of a larger project which, in its entirety, is assessing women s abortion related behavior through, in addition to this component, a morbidity and mortality assessment of abortionrelated complications at a nationally representative sample of health facilities; a community based household survey of women s sexual and reproductive health knowledge and behavior with a special focus on abortion; and an assessment of the capacity of health facilities to perform signal functions. 1

2 Methodology Women who have successfully obtained a legal abortion and women who have experienced complications from unsafe induced abortion were recruited from two public hospitals (one in each district) and a number of private clinics in the two selected districts. For this component, we have employed the use of mixed methods for the data collection: a structured cost questionnaire conducted at the first interview and semi structured in depth interviews (IDIs) conducted at two points in time. Health care providers providing care to the woman were the first point of contact on the study for the participants. Providers were given tables to fill which contained the selection criteria of potential respondents so that they would recruit women according to the study s recruitment criteria. Health care providers caring for women eligible for inclusion introduced the study to the woman and only if she agreed to find out more, was she introduced to the study team interviewer on site. The interviewers were positioned nearby in a private location (a small room or closet in the hospital) in which the interview could be conducted. The Time One (T1) interviews took place at the facility before discharge. 54 IDIs were conducted at T1; in addition, over 100 cost questionnaires were administered to women with induced abortions as well as women seeking care for complications from spontaneous abortion. Three four months after the first interview, the study team was able to successfully follow up with 38 of the IDI respondents and conduct the second interview (T2) to understand the longer term consequences for women of either being able to resume school or work if she had a safe abortion or continued consequences from unsafe abortion complications. T2 interviews took place at a place of the respondent s choosing. Both interviews took place in the language most comfortable for the respondent (English, Nyanja, Bemba or Tonga). Deliberate sampling of induced abortion patients was conducted to achieve a relatively balanced distribution of safe abortion patients getting either medication abortion or manual vacuum aspiration; deliberate sampling also ensured a relatively balanced distribution of patients obtaining treatment for unsafe abortion complications, and of women experiencing low severity complications, and women experiencing moderate/high severity complications. Morbidity was only assessed at T1. For women younger than 16, parental permission was sought. For women ages 16 17, they were treated as emancipated by virtue of the fact that they had been recently pregnant. See Table 1 for a breakdown of the qualitative respondents; we also have quantitative cost surveys for each of them. TOP PAC T1 N (%) T2 N (%) MVA 10 (18.5%) 5 (13.2%) MA 17 (31.5%) 16 (42.1%) Low severity 14 (25.9%) 9 (23.7%) Moderate/severe 13 (24.1%) 8 (21.1%) TOTAL 54 (100.0%) 38 (100.0%) Table 1. Description of Qualitative Sample, Household Consequences of Abortion, Zambia

3 Severity classifications were taken from another component of the parent project, the morbidity component which was assessing the number of near miss cases of abortion related maternal mortality occurring in a nationally representative sample of health facilities (see Appendix 1). This severity classification is a modification of the WHO maternal morbidity severity classification for the Zambian context by Onikepe Owolabi, a study team member and medical doctor, limiting the variables to those captured regularly in women s medical records and/or regularly measured at intake. No doubt for some women, abortion related consequences continued after the T2 interview, especially those who experienced moderate or severe complications. This cut off was chosen to reduce attrition between T1 and T2, and because we felt we could capture the majority of the consequences within this window since with the availability of medication abortion, the severity of abortion complications has reduced in recent years. In addition, had the duration between T1 and T2 been extended, there would have been the possibility that respondents would have forgotten consequences that had happened shortly after the abortion. Furthermore, certain consequences such as infertility and debt would not have been able to be comprehensively assessed at all even within a moderately longer time frame and therefore, we have to accept that we are unable to fully assess these potential consequences. Recruitment was difficult because for the qualitative component, only women who had experienced an induced abortion were eligible to participate. Inclusion relied on women s self reports of induced abortion. Due to the stigmatized nature of abortion in Zambia, and the widely held misperception that abortion is illegal, many women who presented with complications that providers were certain were due to induced abortion insisted to health care providers and interviewers alike their complications were due to spontaneous abortion and therefore were ultimately ineligible for inclusion. As a result, recruitment took longer to complete than anticipated. All women who did participate received 35 Zambian kwacha (ZKW) for their time and 100ZKW for travel reimbursement at each interview (equivalent to ~US$13/interview). Another complication was that health care providers from various facilities in our sample who had been trained and sensitized about the study left their posts for extended periods of time during the recruitment period, either for vacation or further training opportunities. This resulted in gaps in recruitment and a need for the study team to repeatedly train additional health care providers in the study protocol. For each woman recruited into the IDI sample, health care providers were paid a small financial incentive. This was deemed necessary as health care providers are extremely busy and without this incentive, they would have been less willing to make time to help recruit potential respondents. Lastly, even when women would admit to having an induced abortion and they would agree to participate, sometimes others who had accompanied them did not have the time or the interest in allowing the respondent to participate. A quality assurance assessment was undertaken by the Population Council Zambia a few months into the start of fieldwork. A consultant was hired to evaluate the implementation of the recruitment protocol. Her recommendations were extremely helpful to guide the team on course corrections they could implement while fieldwork was on going to improve the processes in place for respondent recruitment and at the various recruitment locations. Course corrections undertaken as a result of her assessment included comparing the providers study recruitment forms with the facility registers at the end of the month to allow us to follow up with providers on cases that they were not referring to our interviewers; changing the hours our interviewers were at facilities to reduce the probability that they would lose the opportunity to speak to potential respondents, and training providers in hospitals in 3

4 other wards (not just maternity wards) to capture women in other wards of the hospital where abortions were also occurring. All IDIs were audio recorded on a digital recorder, transcribed verbatim and then translated into English. The transcripts were then cleaned by members of the study team in Zambia, one of who spoke Nyanja, to be sure the transcripts were written clearly and all local places and names had been stripped from them. The project received approval from the Institutional Review Boards of Population Council Zambia, the Guttmacher Institute (USA), and the University of Zambia Biomedical Research Ethics Committee. Coding of the cleaned transcripts took place in NVIVO 10 (QSR International, Melbourne, Australia), using a predetermined coding structure capturing relevant concepts related to health, economic and social costs borne by the woman related to her decision to obtain a termination. The cost questionnaires were double entered into a database. Using the qualitative case control methodology (Higgins et al. 2014), we matched women on economic status, and compared three groups of women: those who experienced a safe abortion, those who experienced and unsafe abortion which resulted in low severity complications, and those who experienced an unsafe abortion which resulted in moderate or high severity complications. Analyses were conducted on the following themes: sources of expenses and impact on savings. Respondents experiences at T1 and T2 are analyzed as part of a complete narrative for one individual. While not only health, economic and social consequences can change over time, so can women s opinions and attitudes about what she has experienced change over time. Therefore, when describing the results, we identify experiences as well as quotes according to whether they emerged at the T1 or the T2 interview. We present summaries of the findings and illustrative quotes, including the age and the marital status of the respondents to provide context to their quotes. We compare respondents on other demographic variables when salient differences emerged on those dimensions. Analysis is currently ongoing using the framework described above. Preliminary analyses Sources of expenses Women having safe abortions differed in how they paid for their abortions depending on their education level. Less educated women relied on families or partners to raise the funds whereas more educated women paid for the abortion with their personal savings and income. Women who were obtaining PAC tended to be one to three degrees separated from the money that was used to pay for their care. The number of degrees of separation depended on the amount of money needed and the woman s age and personal financial status. Younger, less educated women were most likely to be the furthest separated from the money. Money for their procedures came from family members borrowing money on their behalves. School going women often borrowed money previously allocated for food or school related expenses provided by their families. In contrast, older women either collected the money themselves or borrowed directly from family. Delays in raising the funds increased the cost as well as the risk of the procedure. Women expressed frustration and anxiety about the duration of time during which they were eager to secure an abortion but had to continue to seek funds before they were successful at drawing together enough money to be able to have the procedure done. A 19 year old, unmarried woman obtaining PAC said: But I have been looking for the money which has been so hard to find, which I just found. [The procedure] should [have been] last week. [I was only able to get the procedure] after it had already grown and everything. 4

5 Impact on savings Seeking safe services had a significant impact on household savings. In most cases, the households of women who sought safe abortion were left with no food. In some cases, assembling the costs to pay for the safe abortion forced women to borrow money beyond their ability to pay it back, leaving them in a cycle of debt. Women receiving PAC were often younger and less educated. These women reported knowing less about the impact on the household savings since the money came from other family members. Older women obtaining PAC used their own money or were closely connected to the source of money and reported a significant impact on their savings due to the need to pay for PAC. A 29 year old, unmarried respondent experiencing a severe complication reported at T1: Interviewer (I): So what do you think will be the effects of that now that you have spent so much and that it has left you with no savings? Respondent (R): I will start asking for money now from the relatives[ ] I: Is it going to have any effect on your life, the fact that you have been left with no cash on you? R: Yes, because already there is no power [electricity] at home. There is no food. At T2, she remained in a precarious position: I: What are the consequences that you are going through due to that? Because your spending of the savings that you had is due to the complication that came after the termination right? What s the consequence of using those savings? R: The consequences umm...how can I say it? I will remain with nothing of course. Discussion/Conclusion Unsafe abortion complications cause financial burdens for women which often necessitated that they borrow or use money intended for other purposes. PAC expenses were unplanned expenses, making those expenses harder for the woman and her household to absorb as compared to women who had safe TOP procedures. Younger and less educated women tended to borrow from family members and were less able to describe the financial outlay or the implications of this spending on their household. In comparison, older and more educated women recognized that the abortion often had a significant financial impact on them and their families. In order to reduce the risk that a woman experiences harmful, even catastrophic, financial consequences associated with an unsafe abortion, greater use of family planning and safe abortion services must be promoted in Zambia. Unsafe abortion not only places a woman at greater risk of severe health consequences, but also has a negative economic consequence because of the costs associated with seeking PAC services and the long term health consequences. Acknowledgements We would like to thank UK AID for funding this project; the respondents; the interviewers in Zambia; Onikepe Owolabi for her assistance with the morbidity classification; Anna Abelson and Angel Chelwa, Global Health Corp Fellows at the Population Council Zambia for their assistance in fielding the study, cleaning the transcripts, and contributing to analysis; and Ragnar Anderson, consultant to the Guttmacher Institute, and Meghan Ingrick, Guttmacher Institute, who coded the transcripts. 5

6 References Hendrickson, Cheryl, Tamara Fetters, Stephen Mupeta, Bellington Vwalika, Patrick Djemo, and Keris Raisanen. Forthcoming. Client pharmacy worker interactions regarding medical abortion in Zambia in 2009 and International Journal of Gynecology and Obstetrics. Higgins, Jenny A., Sanyukta Mathur, Neema Nakyanjo, Elizabeth Eckel, Richard Sekamwa, Josephine Namatovo, William Ddaaki, Rosette Namakula, Laura Kelley, Fred Nalugoda, and John S. Santelli The Importance of Relationship Context in HIV Transmission: Results from a Qualitative Case Control Study in Rakai, Uganda. American Journal of Public Health 104(4): The Zambia Legal Information Institute. Chapter 304 of the Laws of Zambia. Termination of Pregnancy Act. act/304. Published

7 Appendix 1. Abortion related morbidity classification criteria for Household Consequences study LOW SEVERITY COMPLICATIONS Temperature < 37.3C (but greater than 36C) No clinical signs of infection No suspicious findings on evacuation (MVA or ERPC or D&C) Haemorrhage/bleeding not requiring any blood transfusion Haemoglobin (Hb) g/dl MODERATE SEVERITY NEAR MISS COMPLICATIONS Temperature >=37.3C Offensive/foul smelling products on examination/mva/erpc/d&c localised peritonitis Haemoglobin (Hb) <9.9g/dl ± Blood transfusion 1 or more units of blood transfused with no laboratory blood test results available Blood transfusion requested for by the doctor in treatment notes, no laboratory test available showing Hb was normal and blood was not given to the patient for any reason Hypovolemic shock Defined as: persistent systolic blood pressure <80 mmhg with a pulse rate of at least 120 beats per minute ± (Hb between g/dl pallor sweaty skin/hands) Septic shock Defined as (a) clinical diagnosis of septicaemia (b) T>39C T<36C genital infection (systolic blood pressure <80mmHg Jaundice an unconscious patient oliguria <100ml in 4h) Hysterectomy Defined as the surgical removal of uterus following infection or haemorrhage/bleeding Cardiac arrest Defined as sudden absence of pulse and loss of consciousness Perforation of the uterus or bowel injury due to an attempted abortion Urine output (<30 ml/h for 4 hours) or (<400 ml/24h) nonresponsive to fluids or diuretics