Table 3.

Iterative responses by discussion group or key informant, highlighting emerged major and minor themes

		Patient couples	Community health workers	Clinical officers/midwives
Definitions	Length of a normal pregnancy	–9-10 months	–36 weeks, which was considered the equivalent of nine months	–40 weeks
	How due date is known	–Just know the month/not the date	–By last period and first antenatal visit	–LMP with gestational wheel
		–The midwife tells you from a wheel	–Most know the month but not the week of LMP	–Ultrasound (“best done at 28 weeks so you can tell due date and presentation of baby”)
		–Only by coming for antenatal care and being seen by a clinician who palpates abdomen	–Not aware of ultrasound for dating a pregnancy
	Earliest age a baby can survive	–6 months if delivered at a health facility (higher chance of baby dying if delivered at home)	–6 months maybe, 7 months yes	–7 months
			–28 weeks (which was estimated to be 2.3 kg)
			–1.5-1.9 kg
		–1 kg	–1 kg if in an incubator followed by Kangaroo Care
		–2 kg
Explanatory models	Causes of preterm delivery	–Beaten by husband	–Stress	–Twins
		–STIs (AIDS, syphilis, gonorrhea, hepatitis)	–Hard work	–Anemia
		–Malnourishment	–Being beaten	–Close spacing of pregnancies
		–Placenta problems	–Malaria	–Overwork
		–Maternal sickness	–Other sickness	–Infections like malaria and STIs
		–Mother too young (<18) or too old (>45)	–Having had many children	–Trauma
			–History of abortions	–Young age
				–Multiparity
				–Incompetent cervix
Access to care	Actions to take if preterm contractions	–Go to health center	–Go to health center (though might not tell family they are going if it is not a scheduled visit)	–Refer to district hospital
		–Go to hospital—which could be 12–20 km away (walk or bike or use ox cart or tractor from nearby estate)	–Might call CHW (20-25% of families have phones)	–Refer to central hospital
	Reasons not to go to health center or hospital	–Church might not condone going to a health center	–Fear of “being cut with sharp things”	–Women and their families may not be able to arrange or afford transport
		–Traditional healer might have herbs to stop contractions	–Fear that nurses and clinicians will be harsh	–“Nothing can be done”
			–Transportation
	Services available at health facilities to treat preterm labor or prevent recurrent preterm birth	–Kangaroo Care	–Rest	–Referral to district hospital
		–Medications to stop contractions	–Cesarean to save mother and baby	–Clinical officer: Not aware of availability of steroids to accelerate fetal lung maturity
		–Never heard of steroids or progesterone	–Referral to district hospital	–Clinical officer: Give salbutemol (terbutaline) and refer if contractions do not subside
			–Never heard of steroids or progesterone	–Referral hospitals offer Kangaroo Care, a nursery, and dexamethasone for mother in preterm labor and cerclage for woman
				–Never heard of progesterone
Openness and acceptability	If progesterone available, route of delivery preference	–Vaginal route (cost of transport to hospital for weekly injection is too high)	–Vaginal route	–Injection (can document adherence to regimen)
	Openness to new concepts	–Would believe that gum disease can cause PTB if given an explanation and the doctor said it was true	–Would believe gum disease can cause PTB “but, first, you have to tell us why”	–“I do not know that gum disease may cause PTB, but if you explain it, maybe I would believe it.”
			–“We would need ‘community sensitization’ first before women will believe this.”
			–Want to make sure previous studies have been done, no harm is being done, and “Malawians are not being used as guinea pigs”
	Social acceptability of new approaches	–Gum or mints are ok	–Gum or mints are ok	–Gum or mints are ok
		–Prefer a jar of gum over a package (can use later for storing other things like salt or medicines)	–Packages are easier to keep in a safe place but jars could be used later for storing sugar	–Have heard of some chewing gum reducing tooth decay
	Optimal delivery sites if preterm labor	–District or central hospital	–District hospital unless free transport is available to central hospital	–District hospital (they will decided if to refer on to central hospital)

To improve trustworthiness, all data was compared and findings were triangulated. All members of the research team including interpreters “debriefed” after each interview or discussion group to enhance the confirm ability of the findings through the representative ranges of perspectives, personalities, and a priori knowledge of the team members.

The ratio of questions asked to answers given were tracked by data recorders. At site 1, 71 questions were asked in total, and 113 answers were given at a ratio of answers to questions of 1.59:1. At site 2, 91 questions were asked and 185 answers given (2.03:1).