Prevalence, care-seeking behaviors and treatment cost for neonatal morbidities in Karnataka (India)


Purpose
Around three million infants die within the first four weeks of life each year – nearly all (98%) of these deaths occur in developing countries. Approximately one million newborns die each year in India. Therefore, this study aims to determine the patterns of reported neonatal morbidity and care-seeking behavior and identify factors associated with it.


Design/methodology/approach
A cross-sectional study was conducted during November 2016. A systematic random sampling technique was used to select the sample. Statistical techniques like Binary Logistic regression and chi-square test were used.


Findings
The results of the study showed that around 31% mothers of neonates reported that their neonates suffered from some kind of morbidity. Fever, jaundice, cough and cold, the low birth-weight and difficulty in breathing were the most common dangers signs reported. Birth order and mother’s knowledge of neonatal danger signs were found to be significantly associated with reporting of neonatal morbidity. In all 95% of the mothers sought care for their newborns. Among those who had problems, 59% consulted private hospitals/clinics, 30% visited District Hospital/Taluka Hospital or higher facilities and another 9% to Primary Health Centers/Community Health Centers. Further, findings show that nearly half of the neonates taken to government facilities have got free treatment, whereas an average cost of 7,156 INR were recorded for treatment, 935 INR for outpatient department and 13,774 INR for inpatient department cases.


Originality/value
There is an urgent need to implement intervention modalities that focus on increasing the level of parental education and access to treatment, and advocating the message regarding newborn danger signs during pregnancy is pinpointed.



Introduction
Neonatal and infant mortality pose a major public health challenge, especially in the developing world, and progress in the reduction in neonatal mortality over the past years has been slow (You et al., 2015). According to the World Health Organization (hereafter WHO) in the year of 2015 around 5.4 million children aged under five years died globally (WHO, 2016;WHO, 2018). Nearly half of all deaths in children under-five occur in the neonatal period, within the first four weeks of life (WHO, 2016;WHO, 2018). Hence, 2.7 million infants die within first four weeks of life each year (You et al., 2015) and nearly all (98 per cent) of these deaths occur in developing countries (Lawn et al., 2005;Dadhich and Paul, 2004). India contributes approximately 25 per cent of all neonatal deaths in the world (Lawn et al., 2005). Although there are no reliable estimates, approximately one million newborns die each year in India. The sluggish decline of infant mortality rates in the past decade in developing countries in general and India in particular could be largely attributed to stagnation in neonatal mortality; deaths in the newborn period now contribute to two-thirds of all deaths under five years (Dadhich and Paul, 2004). It is argued that further decline in infant mortality would require effective implementation of interventions to lower neonatal mortality (Lawn et al., 2005;Dadhich and Paul, 2004).
Infectious diseases such as sepsis, pneumonia, tetanus and diarrhea; preterm birth and complications of birth asphyxia are responsible for most deaths in this period (WHO, 2016;WHO, 2018). Prematurity and intra-uterine growth restrictions are also indirect causes or risk factors for neonatal deaths, especially those because of infection (Lawn et al., 2005). Most of these diseases are readily preventable or treatable with proven, cost-effective and quality-delivered interventions. It is postulated that delayed or inappropriate care seeking behavior contributes significantly to the high neonatal mortality levels in the developing countries including India (Lawn et al., 2005).
Integrated Management of Neonatal and Childhood Illnesses (IMNCI) also envisions that family and community health practices, especially health-seeking behavior, are to be improved to reduce morbidity and mortality in the early years of age (WHO and MOHFW, 2003;Ingle and Malhotra, 2007;Black et al., 2003). But documented literature present a picture of poor health-seeking; in the state of Maharashtra (India), for example, less than 5 per cent of newborns suffering from a major illness were taken to a care provider outside the home for medical care (Bang et al., 2001). Even, though improved a lot in the last two decade, recent round of National Family Health Survey (NFHS-4) says that in Karnataka only around 76.9 per cent children suffering from acute respiratory infection and 69.7 per cent children suffering from Diarrhea were taken to a health facility or provider (IIPS, 2016) Health-seeking behavior is a function not only of the availability of health facilities and other sources of health care but also the motivation and ability of individuals to seek medical treatment (Teerawichitchainan and Phillips, 2007). In this background, the present study is intended to determine the patterns of reported neonatal morbidity and care-seeking behavior and to identify factors associated with it in Karnataka, India.

Methods and materials Ethical consideration
Communication with the District Medical Officer (DHO) and Taluka Medical Officer (THO) was made through formal letter mentioning Ministry of Health and Family Welfare (MoHFW), Government of India's approval to take up the study. Having JHASS 1,2 finished informing the purpose and objective of the study, the researchers obtained oral consent from the study participants with an age greater than 18 years. Participants were informed that their participation was purely on a voluntary base, and the information obtained from them will be kept confidential and will be used only for the research purpose.
Study area and sample size A multi-stage sampling process was adopted to recruit sample from the sub-district. Bidar district[1] of Karnataka[2] state was selected on the bases of maternal and obstetric indicators, and two talukas[3], namely Basavakalyan and Homanabad, were selected purposively among others. Further, five Primary Health Centers (hereafter PHCs) from each block were selected randomly and from each PHC two villages[4] were covered; one headquarters village and one big village in terms of population criteria were adopted, so that a sufficient number of recently delivered women (hereafter RDW) can be obtained. Hence, total 20 villages from 10 PHCs were covered.
Women having a pregnancy outcome during April 2015 to March 2016 were considered as RDW and were identified from the list available with the Auxiliary Nurse Midwife at Sub Health Center (hereafter SHC). A systematic random sampling technique was used to select 460 RDW; 23 RDW were selected randomly from each village. All four hundred and sixty randomly selected mothers were contacted for participation in the study, out of which 376 were successfully interviewed giving a recruitment fraction of 82 per cent. Among them, three mothers had stillbirth; hence, total 373 mothers of neonates who had at least one live birth outcome were included for the neonatal morbidity analysis of the present study. (Figure 1)

Data collection and processing
Pre-tested and face-to-face interview administered questionnaire adopted from different kinds of literature were employed to record mother's reporting about the experience of danger signs among their newborns. The interview questionnaire included separate sections to collect information. In addition to danger signs among  Neonatal morbidities in Karnataka their newborns, socio-demographic, economic and obstetric related information were collected. Seven trained professionals with vast experience in survey research and data collection were conducted the structured interviews, and data was collected during November 14-23, 2016. Completeness and consistency of the data were checked, cleaned and double entered using The Census and Survey Processing System (CSPro) software version 5.0 and analyzed by Statistical Package for Social Sciences (SPSS) software version 20 (IBM Corp, 2011). Frequencies, proportions and summary statistics were used to describe the study population in relation to relevant variables and presented by using tables and graphs.

Outcome variable
The variables used in this study were broadly categorized into outcome variable and predictor variables. The outcome variable includes any neonatal danger sign prevalence. Danger signs are symptoms that complicate the lives of the neonate and happen during the neonatal period, the first 28 days of life. Prompted responses to the total number of 19 neonatal danger signs were recorded and a minimum score of "0" and maximum of "19" was used to measure the experience of neonatal danger signs. Further, two categories were developed for neonatal morbidity, mothers who reported their neonates had at least one danger sign were categorized as "Yes = 1" otherwise "No = 0".

Predictor variables
The demographic and socio-economic variables, such as mother's age, mother's education, father's education, religion, caste, sex of the neonate, birth order of the neonate, number of birth the mother had, knowledge of neonatal danger signs have been used as predictor variables. In this analysis, the predictor variables were categorized as follows: Mother's age: <25 years, 25-29 years and 30 and above years. Mother's education: No schooling, 1-9 years and 10 and above years. Father's education: No schooling, 1-9 years and 10 and above years.

Analytical approach
To assess the determinants of reported neonatal morbidity, first, bivariate analyses were used to examine the nature of association between neonatal morbidity and JHASS 1,2 selected socio economic and demographic characteristics using chi-square test of significance. Second, to examine which factors best explain and predict reporting of any neonatal morbidity among recently born neonates, binary logistic regression model was fitted. For all the statistical tests, p-values of <0.001, <0.01 and <0.05 were considered for statistical significance and odds ratio (hereafter OR) with 95 per cent confidence interval (hereafter CI) was used to identify significant factors. Model fitness test was conducted with Hosmer and Lemeshow goodness of fit test. Table I depicts the percentage distribution of the neonates (0-28 days) born from April 2015 to March 2016 in the study area. Results, as expected, show that the highest proportion of the neonates (around 53 per cent) included in the study were born to the women who were less than 25 years of age and 37 per cent of the neonates were born to the women in the age group 25-29 years, whereas rest of the neonates (11 per cent) were born to 30 and above year old women. Distribution of neonates, by their mother's attainment of education, shows that every fourth neonate was born to the mother who has no schooling or completed less than five years of schooling, and one-fourth of the neonates were born to the mothers who have five to eight years of schooling. Whereas, another one-third and one-fifth of neonates were born to mothers who have completed 9-10 years and more than 10 years of schooling respectively.

Profile of the study population
Around, 79 per cent of neonates included in the study were born to Hindu mothers compared to Muslim[6] (18 per cent) and Christian (2 per cent) mothers. Similarly, majority (42 per cent) of covered neonates were born to mothers belonged to Scheduled Castes (hereafter SCs)/Scheduled Tribes (hereafter STs), whereas, 35 per cent of neonates were born to mothers belonged to Other Backward Castes (hereafter OBCs), and remaining (23 per cent) neonates were born to higher caste mothers. The proportion of male child (55.8 per cent) is more than the female child (44.2 per cent). Further, the proportion of neonates by their birth order reveals that little less than two-thirds of the neonates were of first or second birth order compared to higher birth orders. Seventyone per cent of the neonates belong to below poverty line (hereafter BPL) families and the remaining 29 per cent born in above poverty line (hereafter APL) families. Approximately, little more than two third of neonates were born to the mothers who mentioned at least one key danger sign during neonatal (0-28 days) compared to their counterpart ( Table I).
Prevalence of reported neonatal morbidity Figures 2 and 3 depicts the distribution of neonates included in the analysis had problem or not. Mothers reported the experience of any danger signs in their recently delivered newborns were considered as had problem. Among the total 373 neonatal included in the analysis, nearly one-third (31 per cent) of the neonates were reported to have a morbid condition ( Figure 2).
Further, the result shows that of all mothers who have experienced these signs in their recent delivered newborns, Fever (14.4 per cent) was the most common reported morbidity, followed by Jaundice (7.8 per cent). Six per cent had cough and cold, similarly, low birth weight/Premature (5.6 per cent), breathing difficulty (4.3 per cent) were the most frequent perceived danger signs experienced by mothers ( Figure 3).

Neonatal morbidities in Karnataka
Differentials in neonatal morbidity Prevalence (per 100) of reported neonatal (0-28) morbidity in the study population according to background characteristics is presented in Table II. Information on danger signs that mothers have real experience in their recently delivered newborns was collected in the study, and mothers who mentioned the experience of at least one danger sign is taken as suffering from morbidity. Expectedly, the neonates born to mothers in the early ages say age group of less than 25 years (38 per cent) and later ages, say 30 and above (30 per cent) were more common than the women in age group 25-29 years (22 per cent), and it is statistically significant. Educational attainment of mothers is showing that as the years of schooling increases the reporting of experienced danger signs were also increasing, more or less similar pattern was reported in case of father's educational attainment, but these are not significant.
Reported experience of neonatal danger signs by social and religious community reveals that lower percentage of Muslim mothers reported the experience of any   Similarly, mothers belonged to ST and OBC communities reported any morbidity was 28 per cent and 27 per cent respectively, whereas, a higher percentage of mothers reported any morbidity belonged to general (37 per cent) and SC (35 per cent) groups. Prevalence of neonatal danger signs by biological variables shows that newborn danger signs were more common among male child (35 per cent) than among female child (27 per cent). Birth order of the neonate is significantly evident in reporting any danger signs, as birth order increases the proportion of mothers reporting neonate's morbidity decreases. Neonates of the first order of birth were more common to experience morbidity (45 per cent), followed by second-order birth (29 per cent) and third or higher (23 per cent) birth order. Mother who had knowledge of at least one neonatal danger sign was more common to report neonates experience morbidity than their counterpart and it is statistically significant ( Table II). Table III presents the results of the binary logistic regression model estimated to examine the association between selected maternal, socioeconomic and demographic predictors on reporting of any neonatal morbidity condition. Results reiterate that mother's age is negatively associated with the neonatal morbidity. Women of age 25-29 years (odds ratio [OR] = 0.930, 95 per cent confidence interval [CI] = [0.351, 1.130]) were less likely to report any neonatal morbidity condition compared with women in the age group of less than 25 years. The odds of mothers reporting any neonatal morbidity are increasing for the neonates whose mother's years of schooling increases and it is true in case of the father's educational attainment. Mothers of the female child were less likely to report any morbidity compared with their counterparts (OR = 0.669, 95 per cent CI = [0.415, 1.076]). The odds of neonatal morbidity are decreasing in higher birth order. The knowledge of neonatal danger signs is positively and strongly associated with reporting any neonatal morbidity, mother who had knowledge of at least one danger signs were more likely to report any morbidity (OR = 2.394,95 per cent CI = [1.362,4.209]) compared to their counterpart (Table III). Figure 4 depicts the distribution of neonates for whom medical care was sought for danger signs. Among the total studied neonates, 117 mothers of the neonates reported had a problem, of them 95 per cent (111) sought care for their neonates (Figure 4).

Types of health facilities used
Among those who had problems, majority mothers of neonates (59 per cent) consulted private hospitals/clinics and two per cent took treatment at home. About 30 per cent went to the higher-level government health facilities, such as District Hospitals (hereafter DHs), Taluka Hospitals (hereafter THs) and other higher-level government hospitals. Only 9 per cent visited the lower level government health facilities, such as PHCs and Community Health Centers (hereafter CHCs) (Figure 5), and only 5 per cent have not sought treatment (Figure 4).

Neonatal morbidities in Karnataka
Types of health facilities used by sex of neonates Further, simple cross-tabulation of the type of health facilities used for treatment by sex of the neonates shows that Care sought from the higher level government health facilities was found to be 9 per cent greater for the male neonates than their female Notes: OR = Odds ratio; CI = confidence interval; SCs = scheduled astes; STs = scheduled tribes; OBCs = other backward classes counterparts. Whereas, neonates visited the private hospitals/clinics were found to be 11 per cent greater for female neonates. Such difference was not found among male and female neonates who sought treatment from a lower level of health facilities ( Figure 6).

Cost of treatment by providers
As shown in Table IV, many of the neonate cases going to DH/TH and other higher facilities (42 per cent) and PHCs/CHCs (38 per cent) received free treatment, and some other spent Notes: Percentages sought medical care by type of health provider has been calculated only among those neonate cases who had reported at least one problem and took to health providers; DH = District Hospital; TH = Taluka Hospital; PHC=Primary Health Centres; CHC = Community Health Centres

Sought care 95%
Did not seek care 5% Note: Percentages sought medical care has been calculated only among those neonate cases who had reported at least one problem Neonatal morbidities in Karnataka very nominal amount. Whereas, even though free treatment is available in government health facilities, some of the neonate cases spent a big amount out of pocket, and this proportion is more in higher-level facilities. Whereas, all most all the neonates who visited private hospitals/clinics have spent money out of pocket, moreover, around two-thirds (71 per cent) of them have spent ranging from 1,000 to 70,000 Indian rupees (hereafter INRs), remaining 29 per cent cases have paid up to 1,000 INRs. As the result shows among those who had problems, majority mothers of neonates (59 per cent) consulted private hospitals/clinics, indicating still, with the strong network of government health facilities at various level, many rural mothers are visiting private practitioners and spending out of pocket.
Mean out-of-pocket expenditure Figure 7 and Table V present the average expenditure incurred by the households for neonatal health care services by background characteristics and type of patient care. The analysis of expenditure on these services shows the considerable burden on households, on an average, each household, had spent 7,156 INRs. Further, the results revealed that there are considerable differences among the population groups as well as the type of facility and   (Table V).

Discussion
The present study was intended to determine the patterns of reported neonatal morbidity and care-seeking behavior, and to identify factors associated with the prevalence of reported neonatal morbidities.
Consistent with one of recent study in Bangladesh (Chowdhury et al., 2018), the present study found a high prevalencenearly one-third (31 per cent) of all participated neonates have reported one or the other morbiditythough it was lower compared with findings from previous studies (Awasthi et al., 2006;Ahmed et al., 2001). Fever was found to be the most common symptom which is also truly consistent with earlier studies (Awasthi et al., 2006;Ahmed et al., 2001) followed by jaundice.
Further, the results shows that the majority (nine out of ten) of the mothers in this study reported of receiving care for their sick neonates from health-care providers, which is again consistent with other recent studies (Chowdhury et al., 2018), and much

Neonatal morbidities in Karnataka
better than the earlier studies in the Indian context which is obviously expected (Awasthi et al., 2006).

Strengths and limitations of the study
Present study, though the sample is relatively small, was conducted in rural and hard-toreach areas. However, data in this study were from interviews with an 18-20 months recall period and therefore, were susceptible to recall errors when reporting symptoms and careseeking practice. To address this issue, experienced research staffs were collected the data collection task after receiving rigorously training on appropriate questioning methods and other potential issues related to neonatal danger signs and health-care utilization among others. Further, information on the timing and duration of the symptoms after delivery, and care-seeking were not collected during the survey which restricted us from analyzing and explaining the number of incidence and duration of the morbidities and care-seeking; similarly, the different socio-economic and demographic factors affecting the care-seeking practice was also not analyzed in the present study due to less number of samples. We propose that studies are needed with a relatively large number of samples to address these issues in details.

Conclusion
Over the recent years, with a formulated policy, a lot of money is being pumped by central as well as the state government to empower the health-care facilities/providers to improve neonatal and infant health services at the grass root level, especially in rural areas. By the results of the study, by socio-demographic variables, it is evident that the reporting of neonatal morbidity was more prevalent among young, educated, higher caste mothers. It may be due to the level of awareness of danger signs was more linked to belonging of well off and higher educational attainment of women (Nigatu et al., 2015). On the other hand, the older women less educated and belonging to socially marginalized class respondents were poorly reported the morbidities. Whereas, by obstetric variables, it is mothers of the boy child, first order birth and mother of two or less number of children have reported more neonatal morbidity compared to their counterparts.
Further, interestingly the study found that the majority of mothers were seeking care from private practitioners for their sick neonates, though the strong network of government health facilities is available at various levels, and consequently spending more money out of pocket.
The results of the study suggest that efforts should be made to raise awareness regarding neonatal morbidity, its prevention and the importance of seeking care from trained personnel among expectant mothers, especially mothers belonging to marginalized communities. Therefore, intervention modalities that focus on increasing the level of parental education, access to treatment, and advocating the message regarding newborn danger signs during pregnancy were pinpointed. There is a well-established network of government health-care facilities and medical practitioners along with front line workers at gross root level, they are the major contributors in this regard, hence proper utilization of this government setup is emphasizes. Furthermore, different non-government organizations, private healthcare providers and international organization need to play crucial role. Further, this task can be done using strong Information, Education and Communication materials through using mass media broadcasts including radio, television and even social network forums. While doing so focus should be given for improving the utilization of services in government health facilities, especially PHCs and CHCs, to minimize out of pocket spending.
Notes 1 Districts are local administrative units; they generally for the tier of local government immediately below that of India's sub-national states and territories.