AI Qualitative Analysis for Antibiotic Stewardship

The November 6, 2026 PLoS ONE study found that antibiotic misuse in Madagascar’s poultry sector was driven by knowledge gaps, informal production systems, easy access to medicines, narrow profit margins, and weak veterinary oversight, based on interviews conducted in April 2023. The study demonstrates why AI qualitative analysis can help researchers connect individual interview statements with broader themes across farmers, animal health professionals, and agrovet sellers.

• The study interviewed 79 poultry farmers, 10 animal health professionals, and 10 agrovet shop workers in Analamanga, Madagascar, in April 2023.
• The farmer sample included 13 backyard farmers and 66 semi-intensive farmers, allowing the researchers to compare antibiotic-use patterns across production systems.
• Only around one in ten interviewed farmers had heard of antibiotic resistance, and about half of those respondents either found the concept unclear or believed it applied only to humans.
• Only 4 of the 10 agrovet sellers interviewed had heard of antibiotic resistance, according to the study’s qualitative findings published on November 6, 2026.
• The researchers recommended targeted training, better record keeping, vaccination and deworming schedules, stronger professionalization, and greater support from chick and feed suppliers.

The study found that antibiotic use was embedded in everyday economic and advisory relationships rather than caused by a single knowledge deficit. According to the November 6, 2026 PLoS ONE article, farmers commonly obtained health advice from chick and feed suppliers, veterinarians, agrovet sellers, relatives, or other farmers, while antibiotics were often available without a prescription.

The researchers identified recurring misuse patterns across backyard and semi-intensive farms, including antibiotic use for suspected viral infections, disease prevention, growth promotion, feed changes, and uncertain symptoms. The study also reported underdosing, overdosing, premature treatment discontinuation, and the use of critically important antibiotics such as enrofloxacin and colistin as first-line treatments.

The study found important differences between production systems. Semi-intensive farmers showed somewhat greater knowledge of antibiotic use and resistance, but they also more strongly viewed antibiotics as essential to maintaining poultry health and productivity. Backyard farmers used veterinary medicines less systematically, but they more often relied on agrovet sellers, family members, or other farmers and showed weaker awareness of withdrawal periods.

Economic constraints reinforced these practices. The researchers reported that feed-price volatility, limited access to credit, irregular profitability, and weak bargaining power made preventive investments difficult. One farmer quoted in the study summarized the market pressure this way: "Sometimes you get a lot and sometimes you get a little. The presence of middlemen is also a problem, and the price of chicken feed has become very expensive."

The study also showed why qualitative interviews matter for antibiotic stewardship research. A purely quantitative survey might record whether a farmer used antibiotics, but semi-structured interviews revealed the reasoning, supplier relationships, financial pressures, and informal learning pathways behind that use.

Antibiotic stewardship researchers should analyze stakeholder differences, not only aggregate antibiotic-use frequencies. The Madagascar study shows that farmers, animal health professionals, and agrovet sellers occupy different positions in the supply and advice chain, so a single undifferentiated theme such as "lack of knowledge" would hide important intervention opportunities.

Thematic analysis can translate interview evidence into intervention design. In this study, the strongest intervention targets included chick and feed suppliers who regularly advise semi-intensive farmers, agrovet sellers who influence medicine purchases, and training programs that address vaccination, hygiene, treatment records, withdrawal periods, and antibiotic resistance.

AI-enabled qualitative research can make this comparison more systematic. Researchers can use AI to organize interview transcripts by stakeholder type, production system, location, or topic; identify recurring codes such as prevention, affordability, supplier advice, self-medication, and market pressure; and compare how those codes appear across segments. Human researchers must still validate translations, inspect quotations, review coding decisions, and interpret findings within the study’s sampling and ethical limitations.

The PLoS ONE authors explicitly cautioned that their convenience sample was exploratory rather than statistically representative. The authors recommended representative sampling and rigorous quantitative validation across Madagascar’s poultry sector, so AI-assisted thematic analysis should support, not replace, subsequent epidemiological and quantitative research.

The Madagascar poultry study shows that antibiotic stewardship depends on more than telling farmers to use fewer medicines. According to the November 6, 2026 PLoS ONE research article, effective action must address knowledge, supplier influence, prevention practices, market instability, professionalization, and access to practical training.

AI qualitative analysis can help research teams move faster from multilingual interviews to validated themes, stakeholder comparisons, and intervention priorities. Evidano supports this workflow through transcript analysis, translation, cross-segment comparison, visualizations, and secure AI chat over research documents.

To analyze interviews and open-ended research data with AI-assisted qualitative methods, try Evidano for free.