Caregivers knowledge, practices about childhood diarrhea and pneumonia and their perceptions of lady health worker program; findings from NIGRAAN implementation research project

Background: Despite 60% coverage by Lady Health Worker (LHW) Program, 30% of child deaths in Pakistan are still due to diarrhea and pneumonia. Caregivers are an important stakeholder yet there is little information on their case management practices and utilization of LHW Program. This study explored caregivers’ knowledge and practices about childhood diarrhea and pneumonia and utility of LHW services before and after a supportive supervision intervention. Methods: Cross sectional surveys were conducted with caregivers’ (mothers) pre and post intervention in project NIGRAAN. The intervention aimed to improve LHSs clinical and supervisory skills of lady health supervisors in order to improve LHW performance and ultimately impact caregiver practices. 4250 households were surveyed. Questionnaire was adapted from PDHS 2012-13. Differences between intervention and control groups were assessed using chi square test. P-value of <0.05 was considered significant. Results: Comparing baseline to end line, there were significant overall improvements in caregivers’ knowledge of loose motion (62 to 84%) and dehydration (12 to 18%) as signs and symptoms of childhood diarrhea. There was also a significant overall increase in caregivers’ knowledge of presenting features of pneumoniai.e. fever (58 to 86%), cough (51 to 61%) and breathing problems (25 to 57%). The proportion of caregivers seeking advice for diarrhea from public sector significantly improved in intervention arm from 20% to 29%. Private sector however remained overall preferred choice for care seeking. There was significant overall improvement in awareness about LHWs functioning (93 to 99%) and household visits (91 to 98%). Actual care seeking from LHWs however stayed low (≤ 0.3%) Conclusion: In order to improve utility and expand coverage of LHW Program interventions aimed at providing supportive supervision have the potential to improve caregiver practices and utilization of available services and decrease childhood deaths due to preventable illnesses.


Background
Pneumonia and diarrhea remain the leading childhood killers, together responsible for 29% of all under-five mortality which translates to loss of more than 2 million young lives annually [1]. These deaths occur disproportionately more among the poorest and most disadvantaged children with nearly 90% of deaths occurring in South Asia and sub-Saharan Africa. Pakistan showed a slow progress and under achieved its millennium development goal (MDG-4) targets as the under-five morality fell from 117 deaths in 1990 to 86 deaths per 1,000 live births in 2013 with a gain of only 31 points, lagging behind some of the better performing neighbors like Sri Lanka, Nepal and India [2]. Much of this mortality can be prevented with proven lifesaving interventions for which the global coverage continues to remain low. The Global Action Plan for the Prevention and Control of Pneumonia and Diarrhea (GAPPD) by 2025 identifies that for reducing morbidity and mortality, the appropriate case management of children ill from pneumonia and diarrhea should include improved care seeking and referral, appropriate case management at the health facility and community level, availability of supplies like Oral Rehydration Solution (ORS), zinc, antibiotics and continued feeding during illness [1].
In Pakistan, the community case management practices of caregivers for diarrhea and pneumonia need improvement. Timely recognition of symptoms of diarrhea and pneumonia is of utmost importance by community caregiver as delay in seeking care could be life threatening, especially for pneumonia [3,4]. The recognition of pneumonia by caregivers is generally poor [5]. A recent study in Mirpurkhas Pakistan found that not all caregivers were able to recognize the basic symptoms of pneumonia. Amongst those probed, 40 % did not report fast breathing and chest indrawing as symptoms of pneumonia [6]. Pakistan Demographic Health Survey (PDHS) of 2012-13 reports that only 42% of acute respiratory infections (ARI)

Journal of Community Medicine & Health Education
cases received antibiotics. Of those with diarrhea, 62% did not get ORS and 39% were not taken to a health care provider [7]. Similar results were reported from low income peri urban communities of Karachi with only 41% ORS usage [8]. Moreover, a hospital based study showed that for signs of dehydration, 40% mothers gave nonspecific responses and only 26% responded with sunken eye as the only sign of dehydration [9]. These trends of low recognition of danger signs and care seeking are aligned with studies from other developing countries like Sri Lanka, India, Zambia and Uganda to name a few [10].
The low utilization of public sector health services occurs across all developing countries. A study from India reported that 35% of caregivers consulted private practitioners and 16% went directly to a pharmacy [11]. A recent multicounty analysis of care seeking patterns of parents for their sick children showed that in South Asian countries including Pakistan care seeking was predominantly from the private sector, and community health workers were not an important source of care [12]. An analysis of demographic health survey and multiple indicators cluster survey datasets showed that among countries with high mortality due to childhood acute respiratory illnesses, Pakistan has the lowest (10%) public sector care utilization [13]. Similarly, a study conducted in Karachi on health seeking behaviors of mothers for childhood illnesses showed that only 11.7% sought care from government doctors [14].
The Government of Pakistan's national program for Family Planning and Primary Health Care relies on Lady Health Workers (LHWs) chosen from their community as change agents for bringing health services at doorsteps. Currently this program covers more than half of the country's population including 60-70% of the rural population [15]. Besides other maternal and child health services, the LHW through her limited supply of medicines is able to treat simple illnesses like diarrhea and minor cases of ARI in children under-five.
Despite showing significant improvements in child health services like treatment of childhood diseases, vaccination coverage and promoting breastfeeding [16], the LHWs knowledge and practices for community case management (CCM) have failed to translate to the caregiver level as evident by low recognition and poor case management practices by caregivers for both diarrhea [7][8][9] and pneumonia [5][6][7].
This study explores if there was any improvement brought in caregivers' (mothers') knowledge and practices about childhood diarrhea and pneumonia as well as their perceptions and utility of the Lady Health Worker Program (LHW-P) before and after a supportive supervision intervention.

The project and intervention
With regards to the causes and determinants of stagnant under five mortality rate, the 'Fourth External Evaluation of National Program for Family Planning and Primary Health Care' identified that not only the LHWs lack skills for managing cases of pneumonia and diarrhea in the community but their lady health supervisors (LHSs) also have inadequate supervisory and clinical mentorship skills [17]. To address these shortcomings, NIGRAAN (an Urdu word meaning supervisor) was designed as a randomized controlled trial (RCT) to study the impact of enhancing structured supportive supervision by LHS on LHW's performance. It was hypothesized that ultimately this would translate to improved community caregivers' knowledge and practices of diarrhea and pneumonia. The intervention consisted of training to build LHSs knowledge and skills, clinical mentorship and written feedback to LHWs. For the purpose of NIGRAAN, all the 34 functional LHSs and five LHWs working under each LHS (total 170 LHWs) constituted the study sample. Each LHS served as a cluster. 17 LHSs were randomly allocated to intervention and 17 to control arm [18]. This paper describes the trickle down impact of NIGRAAN's intervention at caregiver level through more structured interactions among LHWs and LHSs.

Study design and setting
Cross sectional household surveys were conducted at baseline (pre intervention in 2014) and then at end line (post intervention in 2015) after an interval of 26 months in District Badin, Sindh.

Sampling technique and sample size
Multistage sampling technique was used to select households with children under-five. The sample size was based on the expected number of children under five who experienced an episode of pneumonia per year. The estimated incidence of pneumonia among children under 5 in Pakistan is 0.41 episodes per child-year (e/cy). In order to achieve 80% power for estimating the effect of the intervention at 95% confidence interval and intra-cluster correlation coefficient of 0.122, it was estimated that in the first stage at least 8,500 households needed to be selected in each arm (one LHW covers approximately 100 households). Second stage of sampling consisted of systematic randomization; taking every fourth household with an expected number of one child under 5 years per household. The final sample to be achieved was thus 2,125 households in each arm [18].

Data collection
The data collection instrument was adapted from PDHS 2012/13. The questionnaire included sections on background characteristics of caregivers, childhood diarrhea and pneumonia as well as the awareness and care giver perceptions of the lady health worker program. The questionnaire was translated into local language (Sindhi), back translated into English to check for consistency in language and then re-translated into Sindhi. The tool was pretested in district Badin prior to survey. The interviews were carried out by trained field workers who were supervised.

Data analysis
The outcome of interest variables pertaining to unprompted care giver knowledge of diarrhea and pneumonia symptoms, practices for care seeking and complimentary feeding as well as awareness, perceptions and utilization of LHW program are reported. The frequencies and proportions are reported for basic demographic and the outcome of interest variables. To assess the pre and post intervention change in knowledge, skills and other parameters between the two study arms chi square test was used. A two-sided P-value <0.05 was considered to be significant for all tests. The data was analyzed using STATA/MP version 13.1.

Ethical approval
The ethical approval was obtained as part of larger RCT from Aga Khan University (AKU) and World Health Organization (WHO). Informed consent was taken from all the child care takers prior to data collection.

Characteristics of respondents
The caregivers who responded to the survey at baseline and end line were quite similar in terms of their socio-demographic characteristics. These include caregivers' age, gender, marital status, education and region. Almost all (98%) caregivers were married females with more than half being 26-35 years of age. Majority (70-75%) had no formal education and belonged to rural areas of District Badin.

Caregivers' knowledge and practices regarding diarrhea
While assessing any change for caregivers' knowledge of diarrhea signs and symptoms, there were significant improvements. Loose motions as a symptom of diarrhea was reported in both intervention and control arms (Table 1). Likewise, the proportion of caregivers mentioning any of the signs of dehydration also significantly increased from baseline to end line (Table 1).  When probed for their health seeking behavior, most preferred private health practitioners. The proportion of caregivers seeking advice from public sector showed improvements in both arms, though significant only in intervention (Table 1). At baseline, only 34% of caregivers in intervention and 33 % in control arm resorted to increasing intake of water during a diarrhea episode. However at end line this practice showed further decline (Table 1). Similarly the proportion of caregivers who increased feeding during illness was very low at baseline (intervention: 4% and control: 3%) and failed to improve by end line.

Caregivers' knowledge and practices regarding ARI
Breathing problem was cited as a key symptom of pneumonia by 25% of caregivers at baseline which increased to 57% by the end line survey. The proportion of caregivers with knowledge of fever and cough as presenting features of pneumonia also significantly increased from baseline to end line (  The private sector was the preferred choice of treatment while seeking care for pneumonia. Use of public sector stayed static in intervention arm while showed significant decline in control arm ( Table 2). Similar to diarrhea, the practice of increasing water intake during an illness episode was low at baseline (23%) and showed a significant decline by end line to around 8%. Correspondingly, the practice of increasing feeding during ARI was actually followed by very few caregivers (3%) at baseline and failed to improve by end line.

Caregivers' knowledge, practices and perceptions of LHW program
At baseline, a vast majority of caregivers were well aware of LHW functioning in their areas but very few considered them capable enough of managing diarrhea and pneumonia in their children under five (Table 3). From baseline to end line, there were significant improvements in awareness about LHWs functioning in the area and their visits to households across both intervention and control areas. However, the perceived LHW capability to provide diarrhea and pneumonia care failed to show any significant differences in both arms ( Table 3). The actual care seeking from LHWs was low (≤ 0.3%) for diarrhea and ARI across both baseline and end line surveys (

Discussion
The study found that comparing baseline to end line survey there were some improvements in the knowledge of caregivers about childhood diarrhea and pneumonia as well as their perceptions about the lady health worker's program. A study by Ciccone et al. Project Leonardo demonstrated that care managers were able to bring about effective improvements in patients' health knowledge, selfmanagement skills and behaviors [19]. However, it is important to consider that the latter study was in a facility setting.
The expected LHW to caregiver interactions focused on diarrhea and pneumonia however did not occur. This is evident by the low utilization rates of LHWs, possibly contributing to lack of expected improvement in CCM practices at the caregiver level. Our results are contrary to a recent review in developing countries which cited around 5% median utility of community health workers for diarrhea and 4% for pneumonia [5]. Moreover, a recent systematic review reported not only improvements in care seeking brought about by community based interventions but also found evidence that the CCM of these illnesses was associated with reduction in associated mortality [20].
Our study results show that despite the high awareness of LHW program and frequent LHW visits, the caregivers lack confidence in community health workers for management of diarrhea and pneumonia in children under five resulting in their low utility as frontline health workers. Several underlying contextual factors need to be taken into account which divert the attention of LHWs from their routine responsibilities. Most notably the focus of program managers and district health officers is also on other activities like supplementary immunization campaigns and family planning. In addition, the ability of LHWs to provide care for childhood illnesses is hampered by several structural issues that have plagued the LHW program in recent years like lack of supplies, transport, poor district health referral system and delayed disbursement of remuneration to name a few [21]. The community caregivers are either unaware of LHW's role in management of diarrhea and pneumonia or bypass them to seek care from other sources, mostly private sector.
While the role of continued drinking and eating during a childhood illness episode is well established, our results show that very few caregivers actually follow this practice. There was a decline in both drinking and eating frequencies between baseline and end line assessments during our intervention that was not anticipated. An RCT from Lahore Pakistan showed that training of community health workers in counseling for complimentary food positively influenced maternal behaviors and practices for supplementary feeding [22]. The LHWs need to counsel mothers to promote complimentary feeding and drinking during an illness episode.
The observation that most of the improvements in caregivers' knowledge in the intervention areas were also transferred to the control areas indicates that in real life, it is difficult to avoid cross over if the intervention and control areas fall within the same geographic locality [23]. In addition, unlike clinical research, the control arm might be affected by other contextual factors. The mere presence of NIGRAAN's supportive supervision intervention with continuous monitoring of LHWs in both arms might have improved the performance in the control group. Moreover, factors intrinsic to the LHW program like the recent regularization of LHWs (before the end line survey) as government employees could have improved their performance.
A study conducted at Haripur Pakistan showed that the care seeking from LHWs for suspected pneumonia increased from 0.7% to 49% following a community mobilization intervention to increase demand from community and availability of medications to the LHWs [24]. Similarly, a study from Matiari district of Pakistan showed that the provision of reimbursement for travelling costs and medications to LHWs resulted in successful diagnosis and treatment of severe cases of pneumonia [25]. Our study failed to show similar changes as a result of low demand from community and lack of essential logistic prerequisites for service provision by LHWs. Trainings focused on childhood illnesses, transportation, medicines and other logistics were some of the contextual barriers hampering LHW performance. Unless these contextual factors are addressed, the utility of this vital community health work force will fail to improve.
The program managers need to seriously consider prioritizing the already mandated role of LHWs in community case management of childhood diarrhea and pneumonia by providing them the enabling environment to contribute to reduction of associated morbidity and mortality.
A general limitation of surveys administered through a questionnaire introduction of measurement or response bias due to errors in the questionnaire design and the data collection process [26]. The way questions in a questionnaire are phrased, translated into local language, asked by data collectors and understood by respondents during administration can also influence results. Therefore, we adapted the questionnaire from a standardized and validated national survey (PDHS) making minor modifications to make it context specific. The whole process of data collection and subsequent analysis was guided by a trained biostatistician who was part of the research team.

Conclusion
This study tried to test whether interactions that were aimed to enhance supervision by LHSs to LHWs could improve caregivers' knowledge and management practices for childhood diarrhea and pneumonia. There were significant improvements in caregivers' knowledge of signs and symptoms of childhood diarrhea and pneumonia in both intervention and control groups. However, specifically in the intervention group, the proportion of caregivers seeking advice for diarrhea from public sector significantly improved. We conclude that the intervention such as training on childhood illnesses, transportation and provision of supplies and medicines can improve LHW performance and contribute to their better utility. A more robust intervention aimed at improving the performance of LHWs and their credibility in the communities they serve is needed. The knowledge gained from this study can serve as a reference for communities with similar situations.