Introduction

The prevalence of low birth weight is around 33% in India¹, the causes of which are multifactorial and knowledge about the risk factors that are modifiable is therefore essential. There have been many studies, which give the importance of risk factors in causing low birth weight individually, but there are only a few studies, which show the cumulative effect of risk factors in causing low birth weight³. The present study was conducted to highlight such cumulative effect of risk factors in causing low birth weight.

Statistical process control tools, popularly called SPC tools^2,5 are known for their simplicity and ease of application in manufacturing processes in industrial production. The present study involves two such SPC tools, the Cause and Effect Diagram (C&E Diagram) and Pareto analysis^2,5. These tools are used to identify and quantify the factors contributing to the effect investigated. The application of these statistical tools in the biological and medical sciences could prove of great value in providing an insight to risk factors contributing most in causing a condition³.

The cause and effect diagram is a visual tool to organize possible causes for a given effect. It is also called Fish-Bone diagram because of its characteristic appearance. The Pareto chart is a frequency distribution of attribute data arranged by category². The Pareto chart is based on the Pareto principle², a simple way of rank ordering the causes by their contribution. Pareto principle² is also known as the 80-20 rule which states that roughly 80% of the effects come from 20% of the causes. These 20% of the causes are called “vital-few” causes and the remaining 80% of the causes are called “Trivial-many” causes. The vital-few causes are the most important causes that need to be addressed on priority. The Pareto chart identifies these most important causes that need to be addressed on priority. Moreover, it is a simple statistical tool, which uses elementary knowledge of mathematics making it easy to use³, at the same time providing information on the critical or vital few factors that should be addressed first.

The present study made use of these simple but effective tools, used extensively in manufacturing processes to bring out the vital-few risk factors associated with low birth weight and show the cumulative effect of these factors in causing low birth weight.

Methodology

Study design:

Subjects

Inclusion criteria

Results

The experts identified 58 risk factors and divided these into six different risk groups as shown in figure 1

15/57(26.31%) risk factors constituted the vital few risk factors and contributed to 76.24% of the total cumulative effect (table1 and fig.2). These are ranked from one to fifteen (Fig 2). The remaining 43/57(73.69%) risk factors contributed only 23.76% of the effect. These constitute the trivial remaining risk factors.

Among the six risk groups categorized, the first 4 risk groups, namely socio demographic risks, medical risks predating pregnancy, medical risks in current pregnancy and environmental and behavioral risks constituted 82.46% of the total cumulative effect (table 2 and fig.3). The other two risk groups namely obstetrical risks in current pregnancy and fetal factors contributed only to 17.54% of the effect.

The vital few risk factors or the risk factors having the greatest cumulative effect in causing low birth weight, identified in this study are all modifiable except sex of the baby. Socio demographic risks, medical risks predating pregnancy and medical risks in current pregnancy contributed to 71.92% of the total effect towards low birth weight (table 2). Table 3 shows the contribution of risk factors and the severity of their risks in causing low birth weight. There are 10 risk factors, 5 each from socio demographic risks and medical risks predating pregnancy and thus have a higher rank order among the risk groups contributing most towards low birth weight. The severity of the risk also is high among the risk factors from these two risk groups.

Table 4 shows the agreement between the prioritization of experts in the brainstorming session with that of the study (both individual and collective). The collective agreement was 47.3% which was statistically highly significant (Spearman’s rank correlation ρ = 0.473, p<0.0001) and individually there was agreement between three experts viz., CM1, OBG1, and OBG2 and the study prioritization to the extent of 75.7%, 44.2% and 31.2% respectively, which was also statistically highly significant (p<0.001) using Spearman’s rank correlation test. There was disagreement between one expert viz., PED2 to the extent of 39.6%, which was also statistically significant (p<0.0001).

Figure 1

Figure1: The Cause And Effect Diagram For Low Birth Weight Babies

Figure 2

Figure-2 Pareto chart showing Rank order of Risk factors contributing towards Low Birth Weight

Figure 3

Figure-3 Pareto chart showing the rank order of risk groups contributing to Low Birth Weight

Figure 4

Table 1: Rank order of Risk factors of Low Birth Weight

Figure 5

Table2: Rank order of risk groups contributing to Low Birth Weight

Figure 6

Table 3: Distribution of vital few risk factors among risk groups

Figure 7

Table-4 Results of Spearman’s Rank Correlation Test for Agreement between the Experts and the study

Discussion

The vital few risk factors identified in this study, indicate the socio demographic status of the population attending these hospitals, which are Bangalore city corporation hospitals, which caters to the urban poor. These risk factors have been found to have a significant association with Low birth weight in other studies, primarily in developing countries. All the vital few risk factors identified in this study are modifiable except the sex of the baby (Table 3). Improvement of socio economic status of the community, improvement of female literacy, ensuring adequate weight gain during pregnancy, consumption of adequate amount of iron and folic acid supplements, delaying the first childbirth will improve the problem of low birth weight. Thus, a collective approach towards the problem of low birth weight will ensure permanent results.

The Cause and Effect diagram and the Pareto chart gives a pictorial representation of both the factors involved in causing low birth weight and the factors which have to be tackled first to prevent low birth weight. The Pareto chart has the advantage of both weighing the risk factors and the frequency of occurrence of such risk factors, so that the intervention measures can be directed towards the “vital few” risk factors, which are both frequent and contributing most in severity rather than focusing on the trivial risk factors associated with low birth weight. These quality control tools not only give comprehensive knowledge on the vital few risk factors but also a logical way to ensure better intersectoral co-ordination towards combating a particular health problem. Vital few risk factors have to be the focus of intervention so that resources can be channeled towards addressing these factors to obtain maximum benefit and permanent improvement of the problem of low birth weight.

The Pareto chart uses simple statistical applications such as mean, cumulative total and percentages, which can be understood easily. With statistically significant collective agreement (table 4; Spearman’s rank correlation ρ = 0.473, p<0.0001)) between the experts and the study results, we can say that brain storming could be an effective tool in bringing out all the possible causes for a particular effect in a short period of time. However, it should be noted that only the experts from tertiary care hospitals were involved. Experts from the tertiary care hospitals tend to provide higher ranking for those causes which they routinely encounter (obstetrical risks) since these are commonly occurring in tertiary care hospitals while the study hospitals might not encounter such causes very often. This could be the reason for the statistically significant disagreement between study prioritization and prioritization by individual experts. This also may be the reason why the obstetric risks in current pregnancy and fetal factors are ranked as the group with trivial risk (table 2 and 3). Employing brainstorming technique to evolve the causes and prioritizing saves time and resources. It also has an inherent disadvantage. The factors that emerge out of the brainstorming exercise are the factors that will be evaluated in the study. Other factors, which may be associated and vital, will not be evaluated even though they may be essential in understanding the risks involved leading to low birth weight babies. It is therefore essential that an appropriate cross-section of experts is enlisted to determine the appropriate risk factors.

The vital few risk factors from this study can in no way be claimed to be universal. It depends on the type of patients attending the health center and the experts in the brain storming session. The C&E diagram and Pareto chart can be tailor made for each level of health care delivery system and appropriate interventions can be designed based on the vital few risk factors that emerge. Thus, Pareto chart and C&E diagram seem to be promising tools, which can be used effectively to plan and evaluate interventions. Now that the statistical tools have led to an acceptable result, a representative population based study on risk factors of low birth weight is required to validate the findings of this study.

Most studies conducted in public health look at the quantum of risk contributed by individual risk factors in causing the effect using relative risks or odds ratios. These might require using multivariate analytical methods like regression techniques which are not easily understood by those who do not have formal training in epidemiology and/or statistics. Many practitioners and primary health center medical officers in India have done undergraduate studies in Medicine and may not be comfortable in doing such analysis. Using C&E diagram and Pareto chart requires elementary knowledge in mathematics and can be used by a large number of such professionals. In addition, once the risk factors or causes are known, the cumulative effect of such risk factors particularly when the disease causation is multifactorial, these statistical process control tools become very useful. Similar studies can be conducted using the C&E diagram and Pareto chart to evaluate factors causing other diseases where the causes are multifactorial in nature e.g. coronary artery disease, accidents, malignancy, naso-bronchial allergy, which are the important diseases in public health.