Study area, design and population
A multicentre cross-sectional study was conducted during the period of one year at the end of 2016 at four Outpatient Therapeutic Programme (OTP) Centre’s in the District Dera Ghazi Khan of Southern Punjab, Pakistan. Dera Ghazi Khan has the worst indicators of malnutrition (stunting and wasting) in Punjab. This socio-economically underprivileged district has the highest prevalence of illiteracy and overcrowding and is frequently affected by floods [16,[i]].
Overall 185 children, aged 6 to 59 months, were selected to participate in the study in case of their parents’ written informed consent. At enrolment, these children had severe acute malnutrition (SAM) without complications according to World Health Organization (WHO) criteria (mid-upper arm circumference [MUAC]<115 mm and a weight-for-height [Z score]<-3), were clinically well, alert and had good appetite. Severe acute malnourished children with complications were not included in the study on baseline assessment. Complications, as defined by WHO, were hypoglycaemia, hypothermia (axillary temperature <35°C), hyperpyrexia (axillary temperature >39°C), anorexia, severe dehydration, grade three pitting oedema, severe anemia and acute lower respiratory tract infection [5].
Sample size
For sample size calculation, WHO Sample Size 2.0 and the following formula have been used:
with a level of significance of 95% (Z=1.96), an expected proportion of wasting (P) of 11% [17], and an expected error (d) of 5%. This leads to a sample size of 151, which was inflated to a total of 185 to enhance study precision, strength, and accuracy. The eligibility to participate in the study was assessed among 252 children; 67 were excluded, because they either did not meet the inclusion criteria (n=52) or parents refused to participate (n=15) (Figure 1).
Data collection and exclusion criteria
A structured and pretested questionnaire was used for face-to-face interviews with parents to assess data on socio-demographic characteristics and nutritional consumption. Interviews with parents and child physical examination were conducted at OTP centers.
Gestational age of mothers was taken from antenatal records in case of hospital delivery or was based on maternal recall if the birth has been conducted at home. The age of children ≤24 months and born before 37 weeks of gestation was adjusted by deducting the number of weeks of missed gestation from the present age.
Medical history was obtained for symptoms of acute diarrhea, high grade fever, lethargy, cough, shortness of breath, seizures, neurological deficit and anorexia. All of these symptoms were indications for a child being clinically ill. In physical examination, severe dehydration, palmar pallor (as an indication for anemia), and severe pitting edema was assessed in children [4, 5]. Children’s vital signs were assessed for hypothermia/hyperpyrexia and respiratory rate. For assessment of hypoglycemia, a heel-prick was done and children with glucose concentration of <3 mmol/L based on the Dextrostix reagent strip were deemed to be hypoglycemic. Appetite test was conducted by giving the child a small amount of ready-to-use therapeutic food (RUTF) to eat. The child who did not consume a minimum one third of a packet (three teaspoons, which is 30g) of RUTF after three tries were labeled as poor appetite [4, 5, 6]. Children evaluated as being clinically ill due to the aforementioned symptoms and classified as having poor appetite were not included in the study. Instead they were referred to an inpatient facility.
Measurements
Anthropometric measurements
Anthropometric assessments were conducted by trained staff that. Assessments were made twice by a study team to avoid potential bias. In case both measurements differed, further assessments were done till a precise value was obtained. The repeated value was then documented. The mid upper arm circumference was assessed to the proximate 0.1 cm with color-labeled MUAC tape at the midpoint among the olecranon and the acromion process.
Child weight was measured with UNISCALE, close to 10 g undressed or in a light dress. If children and infants were not able to stand by themselves, UNISCALE was utilized to determine only mothers’ weight. Afterwards, the infant/child was handed over to the mother while she was standing on the scale and the collective infant and mother weight was assessed. The calculation of infant/child weight was made as the difference among these two readings. UNISCALE was calibrated for standard weight and adjusted to zero prior to all measurements.
The recumbent length of child who were less than or equal to 87 cm in height was assessed to the proximate 0.1 cm with the help of a length-measuring board with a fixed headrest and a mobile foot piece (“SECA GmbH & Co. KG, Hamburg, Germany”), laid on a smooth surface. Children larger than 87 cm in height were assessed in the stand-up position after removing shoes and with heels together on a horizontal flat plate attached to the base of the measuring board [[ii]]. Standard protocols for child growth were applied for calculating the weight-for-height Z-score with WHO ANTHRO version 3.2.2.
Developmental screening
Study participants underwent the Denver Development Screening Tool II (DDST II) [[iii]] for assessing the development profile. This tool measures the children’s proficiency, under the age of six years to do a range of different tasks and make comparisons with a standardized population of same age children. It measures the four parameters of development that is personal-social behavior development, fine motor development, language, and gross motor development. This tool was executed by a physician trained in conducting DDST II for child development assessment. The average screening time for DDST II was about 30 minutes. If the child was uncooperative and untestable at the first screening for global developmental status, a follow up rescreen visit has been conducted after two days. If the child was still untestable, the screening was repeated again after two days. The tool was translated into the native language for elimination of communication barrier. The flowchart of developmental screening is presented in Figure 1.
Following a standardized algorithm, child’s development was categorized as “normal (if child performs the items, on the left of age line completely), caution an intermediate classification (if child is unsuccessful or say no on an item on which age line falls on or between the 75th and 90th percentiles), or delay (if child isn’t passing an item that 90% of children in the standardization passed at an earlier and item on which age line falls completely to the left of the age line)” [19]. These category measurements were then applied to grade global developmental status as “normal” (no category delayed and no more than one category classified as caution), “suspect” (≥2 cautions or ≥1 delay), or “untestable” (based on a specific pattern of refusals) [19].
Statistical analyses
Data was analysed using SPSS version 23. Simple logistic regression was applied to assess the bivariate relationship through unadjusted Odds Ratios (OR) between the dependent and independent variables. The p-value was relaxed up to 0.20 in order to take all significant variables as confounding variables in the multivariate analysis [[iv]]. Hence, multivariate logistic regression was run to measure the dynamics among the potential predictors for the delayed development among children and present results in terms of Adjusted Odds Ratios (AOR) with 95% confidence intervals (95% CI).
The outcome variable was the children’s developmental status as a binary variable (delayed development vs. normal development). Socio-demographic characteristics, medical history and dietary practices are included as independent variables, because they can theoretically affect development delay (Figure 2). The reporting follows the STROBE statement for cross-sectional studies (see Supplementary Appendix 1).