Assessment of Local Supply Chains and Stock Management Practices for Trauma Care Resources in Ghana: A Comparative Cross-Sectional Study CURRENT STATUS: REVIEW

Background : Injuries are a major public health problem globally. With sound planning and organization, essential trauma care can be reliably provided with relatively low-cost equipment and supplies. However, availability of these resources requires an effective and efficient supply chain and good stock management practices. Therefore, we study aimed to assess trauma care resource-related supply management structures and processes at health facilities in Ghana. By doing so, the findings may allow us to identify specific structures and processes that could be improved to facilitate higher quality and more timely care. Methods : Ten hospitals were purposively selected using results from a previously performed national trauma care capacity assessment of hospitals of all levels in Ghana. Five hospitals with low resource availability and 5 hospitals with high resource availability were assessed using the United States Agency for International Development (USAID) Logistics Indicators Assessment Tool and stock ledger review. Data were described and stock management practices were correlated with resource availability. Results : There were differences in stock management practices between low and high resource availability hospitals, including frequency of reporting and audit, number of stock-outs on day of assessment (median 9 vs 2 stock-outs, range 3 – 57 vs 0 – 9 stock-outs, respectively; p=0.05), duration of stock-outs (median 171 vs 8 days, range 51 – 1,268 vs 0 – 182 days, respectively; p=0.02), and fewer of up-to-date stock cards (24 vs 31 up-to-date stock cards, respectively; p=0.07). Stock-outs were common even among low-cost, essential resources (e.g., nasal cannulas and oxygen masks, endotracheal tubes, syringes, sutures, sterile gloves). Increased adherence to stock management guidelines and higher percentage of up-to-date stock cards were correlated with higher trauma resource availability scores. However, the variance in trauma resource availability scores was poorly explained by these individual factors or when analyzed in a multivariate regression model (r 2 =0.72; p value for each covariate between 0.17 – 0.34). Conclusions : Good supply chain and stock management practices are correlated with high trauma care resource availability. The findings from this study demonstrate several opportunities to improve

3 stock management practices, particularly at low resource availability hospitals Background Injuries are a major public health problem globally, incurring more than 11% of all disability-adjusted life years incurred annually.(1, 2) Nearly 90% of this burden falls on low-and middle-income countries (LMICs), which are least prepared to provide trauma care.(3) Opportunely, with sound planning and organization, effective trauma care can be reliably provided with relatively low-cost equipment and supplies. (4)(5)(6) However, availability of these equipment and supplies requires an effective and efficient supply chain. (7)(8)(9) Supply chains include the structures and processes for sourcing of equipment and consumables, purchasing and procurement, transport, and distribution of products. (10,11) The ways in which these structures and processes interact with one another and are organized significantly impacts the availability of essential trauma care resources for patients when needed. (8,12) Examples of systematic trauma care supply chain assessments and targeted improvements in LMICs have not been published. However, other essential services routinely perform and improve their supply chains to strengthen reliable service delivery. (13) As example, family planning resource-related supply chain assessment and improvement interventions in Rwanda reduced the stock-outs of contraceptives through best practice implementation orchestrated by the government and other key partners. (14) Thus, it might be possible to do similar work for essential trauma care in an effort to reduce stockouts of resources required to provide life-and limb-saving care.
A nationwide assessment of trauma care capacity was performed in Ghana in 2015. (8,12) The study also undertook a root cause analysis to identify the factors that contributed to resource nonavailability. The study reported that a lack of supply chain management structures and processes was often responsible for stock-outs of non-drug consumables (e.g. airway supplies, chest tubes, laboratory reagents). (8,12,15) However, that data gathered did not allow specific recommendations to be made regarding ways to improve supply chain management practices.
To address this gap, we study aimed to assess trauma care resource-related supply management structures and processes at health facilities in Ghana. By doing so, the findings may allow us to identify specific structures and processes that could be improved to facilitate higher quality and more timely care.

Setting
Ghana has 16 regions divided into 257 districts. Most districts have several primary health centers (PHC) and a government or mission hospital that serves as a district (first-level) hospital. PHCs provide only basic public health and primary care services; therefore, they were not included in the study. District-level hospitals are staffed by medical officers and nurse anesthetists, usually offer surgical services and have between 50 -100 beds. Injuries requiring more complex care are referred to one of the regional or five teaching hospitals. In addition to medical officers and nurse anesthetists, regional hospitals are staffed by specialist providers (general and orthopedic surgeons) and contain between 100 -400 beds. Surgical services offered at regional hospitals are broader in scope. There are five tertiary care hospitals in Ghana (one of which doubles as a regional hospital); all are affiliated with a medical school or residency program and offer more specialized care.
The public procurement law of Ghana, Act 663 of 2003, as amended, Act 914 of 2016, provides the framework for the health supply chain. Operations experts have noted that this document has inadequate regulations and framework for procurement, transportation, storage and distribution, disposal and reverse logistics of health commodities. (16) Regardless, the law directs health commodities to be managed by a three-tier system: Central Medical Store, regional medical stores, and service delivery points.(9) Ministry of Health coordinates procurement and the Ghana Health Service oversees the operations of the supply chain. (17) These activities are linked to the hospitals through a combined public and contracted transportation system. The Central Medical Store is managed by the Ministry of Health. Each regional medical store is managed by the respective Regional Health Administration. (18) In exceptional cases and only after applying to the Ministry of Health, tertiary hospitals and regional hospitals are allowed to design, organize and execute their own supply chains and work independently with private vendors. They may also procure their commodities directly from the Central Medical Store.

Sample strategy
Ten hospitals were purposively selected using results from a previously performed national trauma care capacity assessment of hospitals of all levels in Ghana. (8) In that assessment, the availability of trauma care resources within 40 hospitals were rated by local stakeholders and direct observation.
The assessment determined that ineffective supply chain management practices significantly contributed to a lack of resource availability to injured adults and children when needed. (8,19) To identify both effective and ineffective supply chain management practices, we selected five hospitals with low resource availability scores for 51 trauma care resources and five hospitals with high resource availability scores for trauma care resources. Trauma care resource availability scores used for this study were the directly derived from the rating scheme of the previous assessment: 0absent; 1 -inadequate, available to less than half of those who need it; 2 -partially adequate, available to more than half, but not to most who need it; 3 -adequate, present and readily available to almost everyone in need and used when needed.(8) Low resource availability was defined as being in the lowest quartile for total trauma care resource availability score (i.e., sum of 51 resource availability scores) and high resource availability score was defined as being in the highest quartile for total resource availability score.
We stratified the 40 hospitals based on their level of care (i.e., tertiary, regional, district/first-level hospital) and their resource availability scores. The 10 hospitals were then sampled at random, assuming they met the following criteria: three district hospitals with low resource availability scores; three district hospitals with high resource availability scores; one regional hospitals with a low resource availability score; one regional hospital with a high resource availability score; one tertiary hospital with low resource availability score; and one tertiary hospital a high resource availability score.

Assessment tools and data collection
Data on supply chain structure and processes of the selected hospitals were collected by using a structured questionnaire adapted from United States Agency for International Development (USAID) Logistics Indicators Assessment Tool (LIAT). (20) The USAID LIAT is the gold-standard for healthcare 6 supply chain assessment in LMICs. We modified the LIAT for the Ghanaian context and reworded some of the prompts to be specific for the trauma care resources described by the previous nationwide trauma care capacity assessment. (8) The LIAT was used to assess the structure and processes of local supply chains at the selected hospitals through face-to-face key informant interviews with all hospital personnel who engage with the local supply chain to minimize the risk of reporting bias by triangulating responses. The LIAT contains 46 questions subdivided into three domains. The domains consist of: i) general stock management (e.g., use of stock system, information technology usage, stock management training, frequency of stock accounting and supervision), supply chain performance (e.g., stock-out rates, duration of stock-out); and iii) storage structures and delivery processes. We compared responses to the LIAT with stock ledgers to further reduce the risk of reporting bias. Six months of stock ledgers were reviewed for number and duration of stock-outs for each trauma care resource. No identifying information was recorded.

Data analysis
Data collected were on paper forms and doubly transcribed to Microsoft Excel v16 (Microsoft Corporation, USA). Differences were rectified using the original form. Supply chain management practices of both groups (five hospitals with resource availability for non-drug consumables and five hospitals with high resource availability scores for non-drug consumables) were described using Stata v14 (StataCorp, USA). LIAT scores from each hospital group were compared and specific successes and challenges were highlighted. The relationship between resource availability and stock management practices was assessed with Chi square test and linear regression. Differences between hospitals were not tested given the purposively small sample.

Ethical considerations
Approval for the study was provided by the ethical committee of the Kwame Nkrumah University of Science and Technology and the University of Washington Institutional Review Board. All data were anonymously recorded. Permission was granted from each Regional Health Directorate and individual hospital administrators (medical superintendents or administrators) before hospital visits.

General stock management
Nearly all hospitals had a stock management record system (e.g., stock card and ledger) that included information regarding stock on hand, quantities used per unit time, losses, and anticipated adjustments (Table 1). Most hospitals verified the ledger and reported the results quarterly. There were minimal differences between low and high resource availability hospitals. Low resource availability hospitals more often had dedicated supply chain management workshops and on-the-job training structures for staff than high resource availability hospitals, which reported less on-the-job training. There were also little differences in the frequency of emergency orders between the hospital groups. Almost all hospitals had flexible and current need responsive methods for determining resupply quantities and used a combination of local contractors and facility staff for product delivery.
Hospitals in both groups most commonly required 2 -4 weeks lead time for ordering, procurement and delivery. The frequency of supervision visitation was not consistent across hospitals and was less frequent in the low resource availability hospital group. Table 2 demonstrates the prevalence of stock management system, including the presence of any and up-to-date stock cards. In general, stock cards were present for most resources. However, low resource availability hospitals lacked up-to-date stock cards compared to high resource availability hospitals. Further, day of assessment stock-outs were markedly less common in high resource availability hospitals.

Stock management performance
The high resource availability hospital group managed a median of 47 of the 51 trauma care resources assessed (range 35 -51 resources) compared to 40 resources at low resource availability hospitals (range 34 -51; p=0.09) ( Table 3). High resource availability hospitals more often had stock cards and up-to-date stock cards compared to low resource hospitals (median 44 vs 35 available stock cards and 31 vs 24 up-to-date stock cards, respectively). High resource availability hospitals were less likely to have a trauma care resource stock-out on the day of assessment (median 1 stockout, range 0 -5) than low resource availability hospitals (median 4 stock-outs, range 2 -14; p=0.05).

Storage conditions
There was no difference in the overall storage conditions between low and high resource availability hospitals with regard to storage conditions (see Table 4; e.g., organized with regard to first to expire, first out; resource damage, protected from temperature extremes, humidity, water and sunlight; insect and rodent control; security; fire safety; overall p=0.55).

Stock-outs
Stock-outs of trauma resources were more common in low resource availability hospitals compared to high resource availability hospitals (39 vs 17 facilities with stock-outs in prior six months, respectively; p=0.01) (Table 5). Similarly, stock-outs tended to be longer in low resource availability hospitals than in high resource availability hospitals over a six-month period before assessment (total of 180 vs 5 days, respectively; p=0.02). Stock-outs were common even among low-cost, essential resources (e.g., nasal cannulas and oxygen masks, endotracheal tubes, syringes, sutures, sterile gloves).

Correlation between resource availability and stock management practices
Increased adherence to storage guidelines and higher percentage of up-to-date stock cards were correlated with higher trauma resource availability scores (Figure 1). Greater numbers of stock-outs and lengths of stock-outs were correlated with lower trauma resource availability scores. However, the variance in trauma resource availability scores was poorly explained by these individual factors or when analyzed in a multivariate regression model (r 2 =0.72; p value for each covariate between 0.17 -0.34).

Discussion
We aimed to specifically assess trauma care resource-related supply management structures and processes at health facilities in Ghana to inform opportunities to improve the availability of life-and limb-saving services. There were several main findings. First, stock management systems were present in all hospitals; however, high resource availability hospitals more often had frequent 9 inspections, up-to-date stock cards, less frequent stock-outs, and shorter stock-outs. This highlights the importance of active stock management practices. Second, there was generally a low-to moderate-adherence with stock storage guidelines, which places essential resources at risk of damage, waste and stock-outs. Lastly, stock-outs of inexpensive essential trauma care resources were commonly reported, particularly at low resource availability hospitals. By highlighting specific supply chain management deficiencies and vulnerabilities, we can better inform the planning and organization of trauma care services.
Stock-outs of essential medicines at the hospital level have been widely reported in sub-Saharan Africa and represent a significant public health challenge with a recognized negative impact on morbidity, mortality and disease epidemiology. (21) Although there are a multitude of possible root causes for stock-outs (e.g., procurement financing and processes, supply capacity, communication and road infrastructure, distribution resources and planning methods, personnel staffing and training, coordination among stakeholders), hospital-level stock management is a common cause of stock-outs and readily addressable. (22) In addition to promoting compliance with existing government, USAID, and World Health Organization guidelines regarding supply chain and stock management best practices, using modeling techniques, training programs, more frequent audits, demand-side incentives, and automated logistic management information systems could markedly improve the availability of trauma resources for injured patients when needed. (21,23) As example, an intervention study in Mozambique that included fifteen hospitals exposed to standard practice or increased frequency of audits, stock management performance reports and incentives for good performance for family planning resources demonstrated fewer and shorter stock-outs in the intervention group.(23) Accountability frameworks built into funding agreements for the management of global public health priority conditions improve stock management at the hospital level. (24,25) Given that trauma care is planned and organized at the national, regional and/or local levels without an accountability framework, specific attention must be paid to the supply chain and stock management practices for essential trauma resources to ensure their availability when needed.
Adherence to stock management guidelines improves resource availability, reduces waste and promotes appropriate resource use. (26,27) Further, provider concerns over stock depletion reduce the use of essential resources when needed. A case study of a tertiary hospital in India demonstrated that improving channels of communication between providers, stock keepers, and an automated vendor management system, increasing the frequency of storage audits, and establishing protocols for stock documentation improved stock management performance indicators (e.g., compliance, documentation, stock-outs).(28) Such quality improvement initiatives are inexpensive and may have a significant impact on resource availability at the hospital-level.
Some of the most common out-of-stock resources were low-cost (e.g., nasal cannulas and oxygen masks, endotracheal tubes, syringes, sutures, sterile gloves). When not due to insufficient funding, stock-outs of low-cost resources are frequently due to a lack of inventory management and procurement processes, which have been widely reported across sub-Saharan Africa. (29,30) Establishing hospital-level low-supply alert mechanisms and real-time catalogs of government medical stores may reduce the number and period of stock-outs. SMS for Life, a public-private partnership that uses text-messaging to flag low-stocks of anti-malarial drugs, was able to reduce stock-outs from 79 to 26% at health centers in rural Tanzania. This study had a number of limitations that should be taken into consideration prior to interpreting the findings. First, the sample size for this study was small. It was our intention to demonstrate the utility of the USAID LIAT for emergency, trauma and surgical resources at the hospital level, was well as identifying specific opportunities to improve the supply chain of hospitals in Ghana more broadly.
We specifically selected hospitals that represented all three levels of care (i.e., tertiary, regional, district/first-level hospitals) and extremes of resource availability to better understand the spectrum of supply chain deficiencies and vulnerabilities. Second, key informants may have over-or underreported in their responses to USAID LIAT questions. To reduce risk of reporting bias, storerooms and stock management ledgers were systematically reviewed to verify the answers given by stakeholders.
Lastly, we did not assess central ordering, procurement, or delivery to hospitals, which can add to the risk, frequency and duration of stock-outs. Although central supply chain management vulnerability and inefficiencies are important, much can be done at the hospital level to improve resource and service availability. Despite these limitations, the findings allow reasonable conclusions to be drawn regarding ways the importance of investigating supply chain management practices and opportunities to strengthen stock management practices at the hospital level in Ghana.

Conclusions
Good supply chain and stock management practices are correlated with high trauma care resource availability. The findings from this study demonstrate several opportunities to improve stock management practices, particularly at low resource availability hospitals. Governments, health systems and hospitals might consider the following recommendations to ensure that trauma care and other essential resources are available when needed: Systematically assess the supply chain and stock management practices to identify opportunities for quality improvement, particularly for resources that lack an accountability framework associated with high priority global health conditions; Incorporate planning exercises, training programs, frequent audits, demand-side incentives, and automated logistic management information systems into hospital stock management programs to improve resource availability; Establish hospital-level low-supply alert mechanisms and real-time catalogs of government medical stores to avoid stock-outs of frequently used resources.
Abbreviations PHC -primary health center LMIC -low-and middle-income country(ies)

LIAT -Logistics Indicators Assessment Tool
Declarations Ethics approval and consent to participate: Approval for the study was provided by the ethical committee of the Kwame Nkrumah University of Science and Technology and the University of Washington Institutional Review Board. All data were anonymously recorded. Permission was granted (i.e., informed assent) from each Regional Health Directorate and individual hospital administrators (medical superintendents or administrators) before hospital visits. Consent for publication: All agencies, stakeholders, and authors have approved this version for