Data collection
The household (HH) survey was initiated only after obtaining consent from the cluster guardians. Lane to lane mapping exercise of the area ensured that all households were covered. During the mapping exercise the number of functional clinics and health posts were recorded irrespective of their registration status. From the PW and RDW who consented and were available on our revisit, information regarding utilization of ANC services and details of services offered during ANC visits were elicited by trained field interviewers using paper based forms. For assessing their dietary intake pattern, a diet survey using a 24 hour dietary recall for 2 consecutive days along with a food frequency questionnaire was conducted by trained nutritionists. Nutritional status was assessed by anthropometry (height and weight) using standard techniques and through biochemical assessment of haemoglobin status using the standard protocol for cyanmethemoglobin method. To gain an estimate of the weight gain trends in the pregnant women, a repeat weight measurement was taken after 2 weeks of the first measurement. Double data entry was done in a database designed in Microsoft Access with inbuilt validation checks. The food frequency was double entered in Microsoft Excel. The dietary recall which was taken as cooked weight from the respondent was converted into actual weights of raw foods consumed. The raw foods were then entered and converted into nutrients using a validated software DIET SOFT version 1.1 (Profound Technical Solutions, New Delhi) which utilizes values from Nutritive Value of Indian Foods [19].