Why was change needed?

The WHO’s 'Field Guide' for the management of MDR-TB, a practical manual intended for health facility workers in low-resource settings, stipulates that reliable contraception – preferably injectable – should be used by all affected women of reproductive age.6 Combined with the rapid upscaling of bedaquiline in the South African National Tuberculosis Programme, this prompted us to investigate rates of contraceptive use in female patients enrolled in our district hospital’s decentralised MDR-TB programme in Estcourt, KwaZulu-Natal. The medical records of all 56 female patients aged 13–50 years commenced on MDR-TB treatment in our programme in 2016 were audited to assess compliance with WHO recommendations. Results are shown in table 1.

We were concerned that low baseline rates of contraceptive use in our population could lead to delay in enrolment into a bedaquiline programme and high rates of pregnancy during MDR-TB treatment, leading to materno-fetal morbidity and increased requirement for termination of pregnancy.

Despite liberal abortion legislation, women seeking termination of pregnancy in South Africa frequently face discrimination and poor access to services. Although at least two-thirds of pregnancies in KwaZulu-Natal are unintended, over 50% of abortions in the province occur outside a designated health facility7 and almost 25% of maternal deaths from septic miscarriage in 2014–2016 occurred as a direct consequence of unsafe abortion.8

We therefore conducted a project that aimed to integrate contraception into the standardised treatment initiation pathway for MDR-TB, increase usage of long-acting injectable contraceptives among women with MDR-TB, and increase the proportion of female MDR-TB patients eligible for bedaquiline therapy. Given that this was a service delivery improvement project in which contraception was provided as part of the routine package of MDR-TB care (as stipulated in both WHO and national treatment guidelines),4 6 ethical approval and patient consent were not considered necessary. All patient data were anonymised during the audit process.

How did we go about implementing change?

During November 2016 we integrated contraceptive services into routine MDR-TB outpatient care by liaising with the hospital pharmacy and procuring a supply of depot-medroxyprogesterone acetate (DMPA) for the MDR-TB unit. Eight specialist MDR-TB nurses were trained in family planning counselling and administration of DMPA by doctors and specialist nurses from the gynaecology department.

Any patient who wished to conceive or who became pregnant during MDR-TB treatment was referred to a tertiary hospital for individualised management.

What outcomes resulted from the change in practice?

A second audit cycle of the same outcome measures in was conducted in all 58 female patients aged 13–50 years commenced on MDR-TB treatment in our programme in the period January–October 2017. Proportions of categorical variables were compared using Pearson’s Chi-squared† and Fisher’s exact‡ tests as appropriate. Results are shown in table 2.

We demonstrated significant increases in: (1) the number of women being commenced on contraception at the time of their MDR-TB treatment initiation and (2) the total proportion of women on long-acting injectable contraceptives by the time they started MDR-TB treatment. The reduction of incident pregnancies from 6.5% in Cycle to 0% in Cycle 2 was encouraging and suggests that these interventions may help to prevent pregnancy in women undergoing MDR-TB treatment. There were no adverse events reported in the women who received DMPA.

Was the change beneficial?

By integrating contraceptive services into MDR-TB care, the proportion of women on injectable contraceptives by the time of MDR-TB treatment initiation increased more than three-fold, from 23.9% to 77.4% (p<0.0001). This represents a significant step towards ensuring that no female MDR-TB patient is excluded from bedaquiline therapy due to lack of reliable contraception.

Limitations

Contraceptive choice in women infected with HIV or at high risk of acquisition is a challenging and rapidly evolving field. We had initially intended to offer the subdermal etonogestrel implant to women commencing contraception at MDR-TB treatment initiation. However, we were deterred by the powerful stigmatisation of this method in the local population: some women had experienced physical assault by their male partners when they discovered the implant, and many believed their antiretroviral therapy (ART) rendered it ineffective. Accumulating evidence supports this: a recent Kenyan study of 24 560 HIV-positive women using contraceptive implants revealed a three-fold higher pregnancy rate in those taking efavirenz-based versus nevirapine-based ART.9 Although our patients had all been switched from efavirenz- to nevirapine-based ART as this is a requirement for bedaquiline eligibility, many retained their beliefs that any ART would invalidate contraceptive efficacy of the implant.

We therefore offered DMPA as the first-line option for women in our study. It had high acceptability to the target population, and concerns that it might increase the risk of HIV acquisition were not substantiated in the recent large, randomised Evidence for Contraceptive Options in HIV Outcomes (ECHO) trial.10

Further limitations include small cohort sizes and the inability to demonstrate sustained improvements across longer time periods. Future work should focus on widening the choice of contraceptives available and addressing the stigma surrounding certain contraceptive methods.

Advice to others considering change

We advise others considering similar intervention to familiarise themselves with attitudes to different contraceptive methods in the target community, to avoid waste of resources and potential patient harm. We appreciate the limitations of this small evaluation but encourage others to embark on similar service integration projects, ensuring that they maintain close collaboration with patients, nurses and managerial teams across community-based and hospital stakeholder sites.