Availability of Essential Commodities and Related Bottlenecks for Community Health System: Systematic Literature Review

Background: This paper explores the extent of community-level stock-out of essential and program commodities for Maternal, Newborn and Child Health (MNCH) among Community Health Workers (CHWs) in Low- and Middle-Income Countries (LMICs), and identies the underlying bottlenecks leading to such stock-outs. Methods: A systematic literature review was conducted of published and grey literature. 48 studies containing information on the number or percentage of CHWs or health facilities (HFs) stocked-out, or reasons for stock-outs at these levels, were included. In addition, interviews were conducted with domain experts from different organizations working on community health to identify primary reasons for stock-outs. The qualitative data regarding reasons for stock-outs were categorized under different themes, using a content analysis method. Results: 48.09% [CI 95%: 39.28 - 56.90] of the Community Health Workers and 54.76 % [CI 95%: 43.54 - 65.98] of the health facilities in SSA countries included in this study experienced stock-outs of essential commodities. A hypothesis test showed no signicant difference in stock-out rates between CHWs and HFs. The most frequently cited reason for CHW stock-outs was a lack of nancial resources, leading to inadequate national level stocks, affecting supply available at the last mile. Moreover, issues at HF and CHW levels in the following areas contributed to stock-outs: transportation, data and estimation of needs, human resources, and stock management and storage. These signicant bottlenecks hinder the ability of CHWs to save lives. Conclusion: Stock-outs of health commodities impact almost half of CHWs, preventing effective service delivery. Many factors contribute to stock-outs, which occur at all of the health supply chain. A system strengthening differing denitions of and inconsistent reporting metrics posed to analyzing results from reviewed studies.


Background
Community Health Workers (CHWs) are paraprofessionals or lay individuals based in the community who provide health services to both urban and rural communities. 1 The role of CHWs is increasingly recognized as a signi cant step towards achieving Universal Health Coverage and the post-2015 health agenda. 2 CHW cadres vary across countries and have different names, levels of training, responsibilities and professional status. 3 They also provide different packages of care in different countries. CHWs play a crucial role in providing (a) access to primary health care services through counselling, and (b) preventive and routine curative health services to families who previously had limited contact with trained health workers 4 in LMICs. 5,6 CHWs signi cantly reduce MNCH morbidities and mortalities 7,8 through distribution and appropriate utilization of health products, however, this effect was only observed if CHWs had continuous access to essential supplies. 9 CHW performance is greatly in uenced by the resources available to these workers. 10 If CHWs are properly trained and equipped, they can reduce child deaths from malaria, pneumonia, and diarrhea by up to 60% through the delivery of Integrated Community Case Management (iCCM). 11 Systematic and narrative reviews have found that stock-out of health commodities is a major hindrance to CHW productivity and motivation. 12,13 Understanding the extent and causes of stock-outs at the community level is critical for optimizing CHW programs. 14

Methods
This systematic literature review documents the extent of stock-outs at CHW and HF levels for MNCH commodities in LMICs and describes the underlying bottlenecks that cause these stock-outs.
We employed two methods: a) a review of published and grey literature on extent of and causes of stockouts at CHW and HF levels; and b) interviews with 30 domain experts, identi ed through snowball method. Stock-outs at HFs were included, as in most settings these facilities are responsible for resupplying products to CHWs. A two-sample test for proportions was applied to determine whether a statistical link existed between the CHW and HF stock-out rates.

Search Strategy
During August and September 2018, we reviewed published and unpublished articles in four electronic databases: 1) PubMed, 2) Global Health via Ovid, 3) Web of Science, 4) Embase via Ovid. In addition, we searched the Google and Google Scholar search engines, and The World Health Organization's website for relevant articles during the same period (see Supplementary File, Appendix 1).
A team of supply chain (SC) experts was engaged to solicit evidence-based, unpublished literature and to identify contact persons in various organizations working with CHWs in LMICs. Grey literature was accessed by requesting documents and reports from identi ed contact persons in various organizations (e.g. USAID, WHO, VillageReach, Save the Children, JSI, MSH, Last Mile Health, World Vision, etc.).
Different platforms, such as the Interagency Supply Chain Group (ISG), Child Health Task Force (CHTF), International Association of Public Health Logisticians (IAPHL), the Community Health Community of Practice, etc. were also approached to identify contact persons.

Inclusion and Exclusion Criteria
In this study, we de ned CHWs as 'workers based in communities (i.e., conducting outreach from their homes and beyond primary health care facilities), providing health services at a community level, who are either paid or volunteer'. 56 de nitions of shortage and stock-out have been used in published and grey literature. 15 For the purpose of this study, we de ned stock-out as: "The complete absence of a required drug at a storage point or delivery point for at least one day." A study was included if it contained a) information on the number or percentage of CHWs or HFs stocked-out, or b) reasons for stock-outs at these levels. The quality of the included studies was assessed by grading, based on the criteria like, geographical coverage, sample size, information on the stock outs, mention of reasons of stock outs and the type of information source -primary or secondary. Study that didn't provide the above-mentioned reasons was excluded.
Considering limited evidence available, we did not exclude any study based on year of study, type of CHW, number of CHWs or HFs assessed, use of primary or secondary data, language, or geographic region. All types of commodities were included, except vaccines, as a large body of evidence exists on the status and root causes of vaccine stock-outs. 16,17

Individual Interviews
A purposive sample of 30 domain experts from 10 countries were identi ed for interviews through the network of contact persons outlined above (see Supplementary File, Appendix 3). Semi-structured interviews were conducted to validate initial ndings from the studies, and to gather additional insights into the underlying root-causes of stock-outs.

Selection and Data Extraction
Once studies were selected, based on the inclusion criteria, relevant data were systematically extracted and tabulated in a Microsoft Excel spreadsheet. The following information was extracted from each study: (1) Study characteristics (including author, title, publication date, year of study, method of study, sample size of CHWs, sample size of HFs, geographic area, types of HFs, and list of commodities), (2) Quanti cation of stock-out (including number and percentage of CHWs or HFs stocked-out, percentage of commodities stocked-out, number of days commodities stocked-out at CHWs or HFs level, and frequency of stock-out at CHW or HF levels), and (3) Reasons for stock-out. We employed the mixed method 18 , in which both qualitative and quantitative information within the same article were collected and analyzed.
The qualitative data regarding the reasons for stock-out were gathered from both studies and domain expert interviews, and then were categorized under different themes, following a content analysis method.

Results
The systematic search of both published and unpublished literature resulted in a total of 1192 records (see Supplementary File, Appendix 2). After the deletion of duplicates, 752 studies were screened in abstracts and titles, resulting in 101 studies. After full-text screening of the 101 studies, 48 studies (including 3 databases) were included in the review; the remaining 53 studies were excluded as illustrated in Figure 1. The included 48 studies contained quanti ed information on stock-out rate at CHW and/or HF levels and/or reasons for stock-outs (see Appendix 4).

Analysis of Stock-out Rates
We computed descriptive statistics of stock-out rates among CHWs and HFs, based on the numbers reported by various studies (See Table 1 We conducted a hypothesis test to understand if a statistical link existed between the CHW and HF stockout rates. The resulting P-value [(0.0734) > 0.05] indicated there was no signi cant difference in the stockout rates between CHWs and HFs (see Supplementary File, Appendix 6).

Percentage of Medicines Stocked-out and Stock-out Duration
Out of the included studies, only seven studies documented the percentage of managed commodities stocked-out among CHWs or HFs. The average percentage of managed commodities stocked-out ranged from 8.4% to 58%. Managed commodities varied by country and included products for iCCM and reproductive health.
Reporting on durations of stock-outs lacked standardization across studies (see Table 2). Studies reported differing periods for stock-outs. This variation is to be expected as the period reported depends on the original data collection method (e.g. through the logistics management information system (LMIS), or review of documents, or through survey/observation on the day of visit).

Reasons for stock-outs
In this review, a bottleneck is de ned as "an obstacle that impedes the availability of essential and program health commodities at the CHWs level".
28 of the 48 studies listed reasons for stock-outs. Semi-structured interviews with domain experts were conducted from October 2018 to December 2018 to gather information on CHW stock-out causes, to identify bottlenecks in SC systems, and to verify ndings from the studies. Information regarding reasons for stock-out were extracted and categorized into different themes. The themes were coded to perform a frequency count. The following SC bottlenecks were mentioned most frequently (Please see Chart 1).

Financial Issues
Budgetary constraint is one of the major barriers to e cient SC functioning, ultimately leading to stockout of community-level commodities. Reviewed studies 19,20 and domain experts also speci ed that inadequate nances for procurement led to national and regional level stock-outs, which reduce community-level stock availability. Lack of nancial resources to procure the entire package of iCCM medicines, particularly non-malaria commodities, was a commonly reported problem. Budgetary constraints also hinder distribution, transportation, supervision, and implementing improved data systems. No details were available on whether nancial constraints were due to the budget development process, allocation of funds, funding ows to subnational level, budget utilization, or a combination of all -highlighting the need for more information.

Transportation
CHWs typically work in rural locations, with di cult terrains. Commuting to resupply points is usually arduous due to poor roads, limited public transportation, or lack of transportation fund. The majority of CHWs in countries, such as Rwanda and Ethiopia 21 , travel by foot for resupply. Lack of funding for transportation to collect supplies from HFs is an impediment limiting product availability. Furthermore, trips to the HF are often linked with monthly supervisor meetings at the HF, which become challenging to hold, for the same reasons.

Data Related Issues
Lack of information systems to capture community-level data, lack of data visibility, and lack of data reporting procedures and tools result in poor inclusion of CHW supply requirements in resupply calculations at all levels. In some cases, MNCH medicines are not even included in the national LMIS.
According to some domain experts, several existing systems do not allow for the disaggregation of stock information between CHW and HF levels. There are exceptions: the cStock system in Malawi provides stock data at the community-level. 22 Good SC data is vital to understand consumption patterns, to inform resupply, to improve quanti cation and management of products.
At HF and CHW levels, poor data collection and use makes it di cult to accurately estimate needs, which are due to lack of capacity among CHWs and poor supervision. Trainings on reporting and data collection processes are limited, as per domain experts, resulting in sparse and low-quality data in those countries that capture data from CHWs.

Human Resource Issues
Human resource challenges include: 1) lack of adequate resources, 2) inadequate staff trainings, and 3) improper remuneration. Inadequate training and limited resources hamper proper data collection and utilization, which often leads to poor demand-forecasting. Although, CHWs have been found to follow procedures and perform simple tasks correctly, given su cient orientation and supervision, most CHWs are either inadequately paid or unpaid, resulting in low motivation. In countries such as Rwanda, more than one-third of CHWs have mentioned remuneration as the major constraint in managing health products. 23

Stock Management and Storage
Limited ability of CHWs and HFs to manage stock e ciently, either due to limited knowledge of stock management practices, or inadequate or improper storage space, also leads to CHW stock-outs. The issue of inadequate storage space is not only con ned to CHWs, who typically store supplies in a box in their home but is more relevant for HFs and regional warehouses. Absence of basic stock management practices and poor storage conditions can affect the quality of medicines or lead to expired or overstocked commodities. 24

Poor Forecasting
Poor quanti cation at the national level leads to inadequate stock levels, which impact CHW stock-outs.
In some countries, CHW requirements are not included in the national forecast. For example, while the cStock system tracks consumption and resupply of CHWs, this information is not considered during the annual national-level quanti cation process.

Issues at Resupply Points
Inadequate supplies at resupply points, absence of or unclear resupply procedures and policies also result in stock-outs. When HFs don't receive enough supplies for their own consumption, they prioritize their needs over CHWs, resulting in CHWs stock-out. 25 Detailed explanations for inadequate supplies at resupply points were not mentioned in the reviewed studies.

Supervision
Effective management of inventory and supply orders require regular supervision. Budgetary constraints often lead to inadequate supervision, due to lack of human resources, capacity building activities, trainings and/or transportation to conduct supervision. Lack of clear guidance also leads to irregular supervisory activities.

Coordination
Communication gaps and weak coordination across SC levels are also the reasons for stock-outs and can often lead to transparency and accountability issues. These issues limit CHWs from knowing whether or when products are available for resupply. 26 In some countries, donor-funded organizations facilitate coordination among SC levels. In certain donor-funded projects, CHWs are restocked during monthly meetings 27 , which are used as a platform to enhance coordination and communication.

Priority/Recognition of CHWs
Some studies and domain experts indicated that CHWs are not fully recognized as a unit of the formal health system (integrated in the national health system). HFs are seen as the "last mile", therefore, CHW stock is, in some cases, viewed as HF stock, rather than separately. Poor recognition of and attitudes towards CHWs from formal health workers, along with the lack of formal policies on the products that CHWs are permitted to distribute, are major challenges in task shifting, despite the fact that some CHWs are experienced and effectively lead medicine distribution. 28

High Consumption and Distribution
Inability to stock adequately during increased demand due to recurrence of epidemics and/or utilization of commodities, also contributes to CHW stock-outs, especially if long time lags exist between the consumption, reporting, and resupply. Varying capacity in distribution at HFs and higher levels and inaccurately captured consumption data at community-level compound this issue. Improper distribution could result in a surplus of commodities at a central level but stock-out at a community-level.

Discussion
While gathering and analyzing data for stock-out of commodities at CHW, we encountered the following challenges:

Limited Study and Inadequate Data
Despite extensive efforts to nd pertinent information, we encountered issues: few studies on CHW product availability have been conducted, the available data on stock-out frequency and underlying causes were poorly documented, and, in many cases, data were inadequate.
Based on our data analysis, most studies have focused primarily on SSA countries. Despite the fact that CHW programs in South Asia and Latin America regions have scored impressive health and social gains 29 , the product availability data from these regions are almost non-existent. Thus, the presented analysis focuses on the SSA region.

Inconsistent Reporting Metrics
Different, non-standardized metrics have been used to report stock availability for community health commodities. Compiling results from various studies is di cult due to such inconsistencies. Standardized benchmarking is, therefore, advised to measure the extent and the causes of such stockouts. Encouraging the use of an existing, standardized set of metrics 30 to report on the stock-out of essential and program health commodities across countries is strongly recommended.
Based on health situations and requirement, different countries have their own lists of essential program commodities that CHWs are allowed to distribute. The reviewed studies were not consistent in terms of reporting such commodities and their stock availability. For example, in some studies the aggregate stock-out percentage of all essential medicines was reported without mentioning the included product types, while in other studies, the stock-out rate of individual medicines was reported.
There is also inconsistency in reported period of stock-outs, which made it di cult to compute or to compare the exact duration. Although some studies mentioned a high frequency of stock-outs, no attempt was made to quantify such frequencies. Information about the total stock-out days along with the frequency are important to understand the extent and impact of stock-outs. These data sparsity and data quality issues impeded us from performing additional informative analyses, such as longitudinal analysis, temporal analysis, and intra/inter-country comparisons. A standard criterion for duration in the de nition of a reported stock-out is recommended for future studies.

Magnitude and Reasons for Stock-outs at CHWs level
Stock-outs present a signi cant issue for quality of care. This paper's results show that almost half [48.09%, CI 95%: 39.28-56.90] of CHWs experience stock-outs of essential program commodities. With the resulting P-value (0.0734) computed for the hypothesis test examining a signi cant difference between stock-out rates at CHWs and HFs levels, we cannot reject the null hypothesis. Based on limited data, it seems that CHWs and HFs experience stock-outs at similar rates, which is not unsurprising as HFs supply the CHWs in the majority of cases.
The most frequently cited reason for stock-outs at the community-level is a lack of nancial resources, followed by other issues. Financial challenge also leads to national-level stock-outs, particularly of nonmalaria iCCM commodities. Besides, issues on transportation, data, human resource, and stockmanagement and storage are signi cant bottlenecks hindering the ability of CHWs to save lives in LMICs.
Given the fact that investment in CHWs is essential to achieve universal health coverage and can result in ten-fold economic returns on investment in SSA 31 , reducing these bottlenecks to promote community health is urgently required.

Strengths and Limitations
We conducted a systematic literature search that included both published and unpublished articles. A number of professional health platforms were utilized to access unpublished literature. We also attempted to gather expert views via personal communications from different organizations working in community health.
Although we employed a thorough approach, it is possible that some publications could have been unintentionally omitted in our review. For example, our query might have missed a number of titles that are used to refer to CHWs. The included studies have used various reporting metrics to present stock-outs at the CHW and HF levels. Because of such inconsistencies, the computed stock-out statistics may not be truly representative.

Conclusions
We attempted to document the extent of community-level stock-out of essential and program commodities for MNCH in LMICs, and to understand the underlying bottlenecks that lead to such stockouts. The result shows that stock-out of health commodities is a major issue affecting almost half of CHWs, thus impacting service delivery and eroding the community's trust in the health system. Stock-outs occur not only at the CHW level but also at higher SC levels. Many bottlenecks that contribute to stock-outs at all the levels of the SC system were identi ed, through reviewed studies and expert interviews. These were primarily related to problems in nancing, data, and capacity to manage stock at HF and CHW levels. Adding the CHW level as an extension to the SC places additional strain on what are often weak supply chains; this structure compounds the identi ed problems. It is recommended to assess whether the SC is robust enough to support iCCM. Several studies have pointed out that more direct SC model reduces the duration and frequency of stock outs compared to conventional three-level system 32 .A system strengthening approach is required to reduce stock-outs at the community-level. Sparse data and inconsistent reporting metrics posed a signi cant challenge for analyzing results from existing studies and resulted in high variance in the ndings. The sparse data on product availability among CHWs warrants additional future studies. Similarly, a set of standard metrics is recommended to measure the rate, period, and frequency of stock-outs in future. * Out of the total included, 26 studies had data on CHW stock-out, 4 studies had data on both CHW and HF stock-outs, and 19 studies had data on HF stock-out rates. Some studies had multi-country information, therefore the number of data points and studies do not match.