IUD uptake remains low in Pakistan, in spite of three major efforts to introduce the IUD since the 1960s, the most recent of these being through the private sector. This study examines barriers to IUD recommendation and provision among private providers in Pakistan.

Methods
A facility-based survey was conducted among randomly selected private providers who were members of the Greenstar network and among similar providers located within 2 Kilometers. In total, 566 providers were interviewed in 54 districts of Pakistan.

Logistic regression analysis was conducted to determine whether correct knowledge regarding the IUD, self-confidence in being able to insert the IUD, attitudes towards suitability of candidates for the IUD and medical safety concerns were influenced by provider type (physician vs. Lady Health Visitor), whether the provider had received clinical training in IUD insertion in the last three years, membership of the Greenstar network and experience in IUD insertion. OLS regression was used to identify predictors of provider productivity (measured by IUD insertions conducted in the month before the survey).

Results
Private providers consider women with children and in their peak reproductive years to be ideal candidates for the IUD. Women below age 19, above age 40 and nulliparous women are not considered suitable IUD candidates. Provider concerns about medical safety, side-effects and client satisfaction associated with the IUD are substantial. Providers' experience in terms of the number of IUDs inserted in their careers, appears to improve knowledge, self-confidence in the ability provide the IUD and to lower age-related attitudinal barriers towards IUD recommendation. Physicians have greater medical safety concerns about the IUD than Lady Health Visitors. Clinical training does not have a consistent positive effect on lowering barriers to IUD recommendation. Membership of the Greenstar network also has little effect on lowering these barriers. Providers' barriers to IUD recommendation significantly lower their monthly IUD insertions.

Conclusions
Technical training interventions do not reduce providers' attitudinal barriers towards IUD provision. Formative research is needed to better understand reasons for the high levels of provider barriers to IUD provision. "Non-training" interventions should be designed to lower these barriers.