The power of positive deviance

Project aim

Setting [Reference]

Positive deviant behaviours

Reduce childhood malnutrition

Vietnam (1999-2000)

Higher consumption of fish, crab, shrimps, snails, breastmilk, clams, beans, peanuts

Boiling drinking water; 5-6 meals/day; active feeding; mosquito bednet; full immunisation; daily bath

Early care seeking for illness

Pakistan (1999)^w13

Higher consumption of sugar, flour, oil, wild vegetables.

Exclusive breastfeeding; intention to breastfeed for 2 years; increased feeding during illness; active feeding; ≥3 meals/day

Greater paternal involvement; supportive family; early care seeking for illness

Improve pregnancy weight gain and birth weight

Egypt (2000)¹³

Higher consumption of meats and vegetables

More antenatal visits; more daytime rest

Care-seeking to decrease painful urination or discolored urine

Improve household maternal and newborn care

Pakistan (2001)¹²

Saving funds and obtaining transport for emergency; tetanus vaccination

Assuring clean delivery surface, attendant’s hands, and blade

Receiving blanket; leaving cord stump undressed

Exclusive breastfeeding

Danger sign recognition with prompt care-seeking

Reduce girl trafficking

Indonesia (2004)

Fear of losing contact with daughter, of disease, or of a bad experience; shame from sex work or violating religious tenets

Planting multiple crops (beyond coffee); reducing family expenses so daughter does not have to work outside village; establishing daughter in small local business

Discussing risks of entertainment industry work; identifying and avoiding neighborhood brokers; monitoring daughter’s friends; reiterating family values; investigating out-of-village work possibilities; requiring daughter to report home regularly

Design and study setting

Results*

Retrospective analysis of project weights (n=1893); Thanh Hoa Province, 1993-1995

Nutritional status of all children improved (d-WAZ +0.36, from - 2.14 to - 1.78, P<0.001). Nutritional status of severely malnourished children (<−3 Z) improved dramatically (d-WAZ +1.44, from -3.58 to -2.14, p<0.001)⁵

Comparison of former project children and their younger siblings with age and sex matched sibling pairs in non-intervention commune (n=71), three years after project stopped; Thanh Hoa Province, 1998, 1999

Older former program children were somewhat better nourished than their non-program counterparts; and their younger siblings, born after the intervention ceased, were much better nourished than their non-program counterparts (WAZ -1.82 vs. -2.45, P=0.007; WHZ -0.71 vs. -1.45, p<0.001).¹⁶ Former program mothers fed their younger children more than non-program mothers (2.9 vs. 2.2 meals/day, p<0.001).

Prospective, randomised, evaluation of the effectiveness of a positive deviance informed integrated nutrition project. Weight measured monthly for first 7 months and at 12 months. (n=240); Phu Tho Province, 1999-2000

Younger (<15 months of age) malnourished (WAZ <-2) children at baseline (n=35) had less deterioration than comparison children (d-WAZ in first four months -0.05 vs. -0.20, p=0.02; and d-HAZ over 12 months -0.10 vs. -0.23, p=0.01).¹⁸ Intervention children consumed intervention-promoted foods more frequently (PD food/day 4.1 vs. 3.6, p<0.05), ate more food (410 vs. 340g, p<0.01), consumed more energy 827 vs. 718 kcal/d, p<0.05), and were more likely to meet their daily energy requirements (49 vs. 35%, p<0.01) than comparison children.^w14 Intervention children experienced less respiratory illness than comparison children (adjusted odds ratio 0.5, p=0.001).¹⁹