Knowledge on complementary feeding among parents of children aged 4-12 months attending a base hospital in a rural district in Sri Lanka

Introduction: In developing countries, inadequate knowledge on complementary feeding (CF) is one of the main reasons for malnutrition Objective: To assess the knowledge on CF among parents of children aged 4 to 12 months attending a base hospital in a rural district in Sri Lanka. Method: A descriptive cross sectional study was conducted in the paediatric ward and clinic of Base Hospital Bibile from April to June, 2016. A total of 383 parents of babies aged 4-12 months were selected by consecutive sampling. Their knowledge regarding CF was assessed by using a structured, pre-tested, interviewer-administered questionnaire. Questions were classified as ‘must know’, ‘better to know’ and ‘good to know’. Results: The response rate was 100%. Mean total knowledge score was 33.79 ± 4.3 (total 49) and the distribution was positively skewed. Mean knowledge scores were 7.08 ± 1.95 (total 11), 9.67 ± 1.88 (total 15), 17.03 ± 2.08 (total 23) in ‘must know’, ‘better to know’ and ‘good to know’ sections respectively. A significant correlation was only observed between age of the child and the total knowledge score (p<0.05). Only ethnicity (p<0.005) and religion (p<0.05) were significantly associated with total knowledge scores. Multiple logistic regression revealed only family income as a predictor for satisfactory knowledge among parents. Conclusions: Parents of children aged 4 to 12 months attending Base Hospital, Bibile had satisfactory knowledge on CF. DOI: http://dx.doi.org/10.4038/sljch.v46i2.8270 _________________________________________ Acting Consultant Paediatrician, Base Hospital Bibile, Consultant Paediatrician, Lady Ridgeway Hospital for Children,Colombo *Correspondence: viduseram@yahoo.com (Received on 18 August 2016: Accepted after revision on 23 September 2016) The authors declare that there are no conflicts of interest Personal funding was used for the project. Open Access Article published under the Creative Commons Attribution CC-BY License (


Introduction
Introduction of safe and nutritious foods at about 6 months of age in addition to breastfeeding (BF) is referred to as complementary feeding (CF) 1 . According to the World Health Organization (WHO), CF should be given timely, adequately and appropriately 2 . In developing countries, inadequate knowledge on CF is one of the main reasons for malnutrition 3 . Understanding the prevailing feeding practices of infants and young children is crucial for implementing or improving health and nutrition programmes in a country 4 . Some common feeding malpractices are early or delayed CF, offering foods with unsatisfactory energy and nutrient density, giving small amounts of food with thin consistency and cultural beliefs contributing to restrict or avoid some nutritious food 5 . The current study was undertaken in Bibile to find out the knowledge of parents on CF as such studies have not been carried out previously in the Moneragala District.

Objectives
To assess the knowledge on CF among parents of children aged 4 to 12 months attending a base hospital in a rural district in Sri Lanka.

Method
A hospital based descriptive cross-sectional study was carried out in the paediatric ward and paediatric clinic of Base Hospital (BH), Bibile from 1 st April 2016 to 1st July 2016, a period of 3 months. The study population comprised the parents of 4-12 month old children admitted to the paediatric ward or attended paediatric clinic in BH, Bibile. Parents of children with known anomalies, cerebral palsy, malignant disorders or other chronic conditions, parents whose child was very sick needing emergency care, parents who failed to give consent for the study and parents whose child was on a special diet or on tube feeding were excluded.
Study sample: Consecutive sampling was done. All admissions to the paediatric ward and all who attended the paediatric clinic during the study period were included. The following formula was used to calculate the sample size 10 : n = Z² p (100p) / d² where n = calculated sample size, Z = Z value corresponding to the required level of confidence (1.96), P = level of knowledge of parents on CF (50%) and d = required level of precision (7.5%) n = Z² p (100-p) / d² n = 1.96² * 0.5 * 0.5 / 0.075² n = 171 Design effect was taken as 2 Final sample size = 171 x 2 = 342 Data collection: A structured, pre-tested, interviewer-administered questionnaire was used. The questionnaire was translated into Sinhala and Tamil by a medical person and re-translated into English by a non-medical person. Differences between the original English version and the translated version were sorted out by discussion. Variables in questionnaire included general information of child, general information of parent, knowledge on CF and source of information on CF. Two bilingual health volunteers were recruited as data collectors. Principal investigator also did data collection whenever possible. Half a day training and half a day practical experience were given by the principal investigator before the study proper. They were given a brief training in research, survey procedures, and ethics, including confidentiality in research prior to data collection. Each interview took 10-15 minutes and was based on the questionnaire which the data collector filled out simultaneously. The interview was conducted in a prearranged room in BH, Bibile at a time convenient to the parent. Only one parent was selected for the interview.
The questionnaire contained demographic and other relevant information. There were 49 questions to assess knowledge on CF. Once the interviewer completed the questionnaire, it was rechecked to determine whether all questions had been marked. The questionnaire was pretested with a sample of 25 eligible parents attending BH, Bibile. During the pre-test, special attention was paid to understanding the questionnaire, cultural acceptability, clarity, simplicity, response rate, quality of responses, and the time taken for completion. Following the pre-test a few modifications were made to the questionnaire. Each question was worded clearly and unambiguously to ensure uniformity. Jargon was avoided and questions worded in such a way that the "respondents" could understand and respond without difficulty. Confidentially was assured both verbally and in writing. A request was made in writing to the respondents to answer the questions truthfully. Re-administration of the questionnaire was done for non-respondents. Reliability was assessed by test re-test reliability method. Leading and presuming questions were avoided.
Data analysis: Data entry was done using epi-data software. Data analysis was done using SPSS 20.0. Percentage, proportions and contingency tables were used for description of data. Association of knowledge of CF with socio-demographic characteristics was analysed using the Chi-square test. A p value <0.05 was considered significant. Frequency distribution tables were prepared to describe knowledge on CF and sources of information gathered. The difference between two proportions was used for binary variables and Spearman r correlation coefficient was done to assess the significance against basic sociodemographic characteristics. There were 49 questions to assess their knowledge on CF of which 11 were identified as 'must know', 15 as 'better to know' and 23 as 'good to know'. Each correct answer was given 1 and each incorrect or don't know answer 0. Therefore, maximum possible mark was 49. Level of knowledge was categorized into four groups, 'highly satisfactory' when more than 75% knew correct answer, 'satisfactory' when 50-75% knew correct answer, 'not satisfactory' when 25-49% knew correct answer and 'highly unsatisfactory' when less than 25% knew correct answer.
Ethical issues: Ethical approval for the study was granted by the Ethics Review Committee of the Sri Lanka College of Paediatricians. Informed consent was obtained from each participating parent. They were given an opportunity to ask questions. All data collected were kept with the main investigator and used only for analysing outcome. There were no questions which disclosed any sensitive or confidential facts. This study did not involve any invasive or painful procedures. Permission to conduct the study was obtained from the Medical Superintendent, BH, Bibile and Regional Director of Health Services (RDHS), Moneragala.

Total 383 (100)
Fifty six percent of fathers and 62% of mothers studied up to Grade 10. Only 0.3% in each group had a university or higher education. Nearly 1% among fathers had no schooling.
Of the 383 main caregivers, 16 (4.2%) were employed and 367 (95.8%) were unemployed. In 215 (56.1%) instances extended family support was available at home whereas in 168 (43.9%) instances there was no extended family support. Table 2 shows the type of occupation of the parents in the study sample. Ninety six percent of the mothers were house wives whereas 39% of the fathers were skilled agricultural and fishery workers. There were no professionals among the mothers while only 0.5% of the fathers were professionals.  Analysis of the knowledge of main caregivers on CF is shown in Table 3. The responses to the question "What do you think about adding some butter / oil for your child's meal?" are shown in Table 4. Eighty six percent of main caregivers knew that oil will help to increase the calorie content. Thirty eight percent of main caregivers knew that oil will not adversely affect the child by increasing the cholesterol level. Ninety two percent of main caregivers knew that oil will enhance palatability and 94% knew that oil will increase weight gain of the child. Oil is recommended for infants less than 9 months of age and 73.6% main caregivers knew this fact correctly.
The responses to the question "If the child refuses to take his/her meal, what are the proper ways of persuading him/her to eat?" are shown in Table 5. Ninety nine percent main caregivers knew that child should be talked to kindly and lovingly, 98% knew that scolding the child is a wrong practice, 98% knew that frightening/ threatening a child is a wrong practice and 95% knew that using physical restraint and forced feeding is wrong. Offering a variety of food with different tastes and textures is recommended to persuade the children to eat and 91% main caregivers knew this fact. Storytelling and taking the child out and showing birds are not recommended to persuade children to eat. Only 5.7% and 3% of the main caregivers respectively knew the correct answer. Allowing children to watch TV is not recommended as a method of persuading a child to eat and 70.5% of the main caregivers knew this.
The responses to the question "What should be the consistency of food which you offer at 6, 8 and 12 months of age?" are shown in Table 6.
Ideally smooth paste is recommended at 6 months of age, smashed with pieces at 8 months of age and whole grain/pieces at 12 months of age. More than 85% of the main caregivers knew the correct answer to each question.   Table 8. The responses to the question "Why do you think it important to be knowledgeable on the proper way of introducing complementary food to your child?" are shown in Table 9. Ninety one percent knew correctly that it will help to minimize the faltering of weight with weaning. Only 44.4% of the respondents knew that it will help to reduce iron deficiency anaemia in children.
Only 42.6% of the respondents correctly answered that it will help to prevent stunting. Nearly 90% of the respondents knew that knowledge on proper way of introducing complementary food is important to optimize your child's mental development and school achievements.
The responses to the question "When you are preparing complementary food, are you influenced by anyone?" are shown in Table 10. The majority said they were not influenced. The responses to the question "How do you gather information on complementary feeding?" are shown in Table 11. The most common source was the Child Health Development Record (62%) followed by the Midwife (28.7%). Only 1.3% of the respondents said they gather information from hospital staff.
Analysis of the level of knowledge in the study sample is shown in Table 12.   The mean total knowledge score was 33.79±4.26 and the distribution was positively skewed. Level of knowledge among parents was classified into 4 levels, highly satisfactory (>75%), satisfactory (50-75%), unsatisfactory (25-49%) and highly unsatisfactory (<25%) depending on the mean knowledge scores in each category of questions.
Accordingly, each category showed satisfactory results. The total knowledge score was 69%, 'must know' category 63%, 'better to know' category 66% and 'good to know' category 73%.
Correlation between knowledge scores and selected socio demographic data is shown in Table 13. Relationship between knowledge scores and selected socio demographic data is shown in Table  14. Only ethnicity (p<0.005) and religion (p<0.05) were significantly associated with total knowledge scores.

Multivariate analysis of correlates
Direct logistic regression was performed to assess the impact of 4 independent variables (age, family income, religion and ethnicity) on the likelihood that respondents would report that they had satisfactory knowledge. Two independent variables, family income (p<0.005) and ethnicity (\ p<0.05) were significantly correlated with satisfactory knowledge. The strongest predictor of satisfactory knowledge was the family income.

Discussion
Most participants in the current study were Sinhalese Buddhists reflecting the ethnic and religious composition of the District. A hospital based study done in India in 2006 found that 46% of mothers had correct knowledge about recommended time for starting CF 6 . In our study 92% of main caregivers had correct knowledge regarding initiation of CF. Differences between the 2 studies may be due to the well-established community midwife service and high literacy rate in Sri Lanka.
Knowledge of main caregivers was 'highly satisfactory' (>75%) regarding the correct times to introduce water, fruit, vegetables, fish and salt to the child's diet, correct time a child should be able to eat from the usual family diet, the correct time to introduce finger food to their child, the correct consistency of food to be given to infants at 6, 8 and 12 months of age, the best way of feeding a child after 9 months of age, the number of main meals and the number of snacks that should be given per day by 9 months of age and the best position to feed a child. Knowledge of main caregivers was 'satisfactory' (50-75%) regarding the correct times to introduce meat, egg yolk and sugar to the child's diet, the correct time a child should be fed and the best environment to feed a child. Knowledge of main caregivers was 'unsatisfactory' (25-49%) regarding giving the child feeds after 6:00 pm when necessary and increasing the number of meals being offered when the child is ill.
Only 38% of parents knew that oil will not adversely affect the child by increasing the cholesterol level. It is recommended to continue with complementary food during an illness to maintain an adequate nutrient intake, which is vital for speedy recovery 7 . In our study, majority of parents had poor knowledge on correct feeding practices while child is having diarrhoea. Only 37% knew that child should be given the usual food depending on the child's appetite. A previous study in Sri Lanka reported similar results 8 . Only 20-24% of parents were aware that 'on demand' breast feeding and regular and frequent breastfeeding should be halted once CF has started. Only around 40% of parents were aware that CF will help reduce iron deficiency anaemia and help prevent stunting in children.
The current study used an interviewer-administered questionnaire which greatly reduced the number of unanswered questions compared to a selfadministered questionnaire. The total knowledge score was 69%, the 'must know' category 63%, 'better to know' category 66% and 'good to know' category 73%. A significant correlation was only observed between the age of the child and the total knowledge score (p<0.05). Only ethnicity (p<0.005) and religion (p<0.05) were significantly associated with total knowledge scores.
There is a consistently noted link between proper infant feeding and the educational level of parents 9 . Mother's education was found to have a significant impact on appropriate infant feeding after a comparative study involving 5 Asian countries 10 . The present study did not find any association between parental education level and knowledge on CF. The present study revealed that the majority of the parents were not influenced by anyone when preparing complementary foods. The most common source of information on CF was the Child Health Development Record followed by the Midwife. This was a hospital based study and may not have represented the community as whole. It is possible to introduce recall bias in collecting information regarding feeding knowledge in retrospect. The bias is likely to be higher when a parent has more children.

Conclusions
Parents of children aged 4 to 12 months attending Base Hospital, Bibile had satisfactory knowledge on CF.