Table 3 Implications of key formative research findings for intervention design

Insights from contraception literature

• Health concerns identified as major reason for non-use. Other reasons include factors related to access, cost, autonomy

• Intervention needs to address health concerns as well as factors related to access, cost (by nforming clients where they can access contraception near their home) and autonomy

• Limited evidence for interventions to improve adherence to specific contraceptive methods or uptake of PAFP

• The intervention needs to anticipate some discontinuation and aim to facilitate safe method switching and well as support continuation with existing method

• Most discontinuation occurs within the first few months

• Decided to provide intervention for three-months

Insights from mHealth intervention and behaviour change literature

•Uni-facteted* adherence interventions have at best modest effects

• Developed a multi-faceted intervention providing information reminders and support to boost motivation to use PAFP

• A semi-automated mHealth intervention increased adherence to HIV treatment in Kenya

• A similar intervention could be adapted for PAFP in Cambodia

Case note review

• 40 % uptake of effective PAFP at the time of seeking abortion services

• An mHealth intervention is an opportunity to maintain contact with clients that don’t return to the clinic for contraception after seeking abortion services

•Over 50 % clients did not return to the clinic within 12-months

Interviews

• Side-effects with contraception common

• Re-enforced findings from literature that intervention should address health concerns

• Clients can find it difficult to make decisions about PAFP at time of seeking abortion services

• The mHealth intervention is an opportunity to maintain contact and remind clients about available methods

• Women sometimes have to discuss with their husband/partner before using contraception

• Re-enforced findings from literature review that the intervention take into account women’s lack of autonomy, facilitating a discussion with husband/partner if appropriate

Focus group discussions

• Preference for voice rather than text-based intervention

• Intervention used voice messages sent to clients phone instead of text-messages

• Many clients preferred direct phone call to automated message

• Developed a semi-automated intervention as fully counsellor delivered intervention would be costly to scale up

• Clients preferred that the messages mentioned the terms ‘Marie Stopes’ and ‘contraception’

• Voice message mentioned ‘contraception’ and ‘Marie Stopes’, but not the name of the client

Consultation with MSIC staff and other organisations

• Text-message interventions likely to have limited success in Cambodia

• Re-enforced findings from clients that intervention should use voice rather than text

• A fully counsellor delivered intervention would be costly and hence harder to scale-up

• Intervention was semi-automated aiming to identify clients most in need of additional support