Questions:

1. Interpretation of Frac_Time_Used_By_OfficerType in the healthsystem logger. I've currently assumed that the values give the total number of staff used in the simulation.
2. Is it correct to ignore the squeeze factor because the actual capability used is costed? The same costing formula should work for modes 1 and 2.
3. How is Frac_Time_Used_By_OfficerType recorded when capabilities_coefficient > 1 --> should costs use capabilities_today?
4. The costing is currently disaggregated by Facility_Level and Officer_Type. For the calculation of intervention ICERs, we will need to calculate cost by Treatment_ID. Does the healthsystem logger capture health worker time by Treatment_ID?

Draft figures for HR salary cost by cadre and level =-

I have now created two categories of costs - 1. Financial Cost and 2. Economic Cost. Here is how I envison HR costs will look under the two categories -

Currently the financial cost of HR does not match well with the 2018 salary budget -

Consumables costing
The cost of consumables during a year is not only what is actually dispensed (as recorded in a simulation) but the cost of purchasing the stock which is needed to dispense that quantity. Therefore we need to multiply the consumables required in the simulation with a ratio > 1 to estimate the cost. This figure explains how this can be estimated in theory (slightly roundabout but I found it helpful for my own reasoning). Based on accounting principles, the consumable stock relevant for this year's costing is not only what is purchased this year but also what is transferred from the previous year's purchase and less what is transferred to the next year. Any expired (or stolen) quantity out of this net inflow is also a cost for this year.

Possible ways to derive x in practice -

1. Use OpenLMIS data to understand the ratio of inflows to outflows. But data for several months is missing.
2. Use a literature based estimate

References for drug theft and expiry:

1. Kebede et al (2021) - Storage management and wastage of reproductive health medicines and associated challenges in west Wollega zone of Ethiopia

• facility-based cross-sectional study from 15th to 31st July 2019 using quantitative and qualitative data from West Wollega Zone of Ethiopia.
• The total value of reproductive health medicines wasted due to expire in surveyed facilities was 357,920.52 ETB (12,323.81 US dollars) and the Percentage of Stock Wasted due to Expiration was 8.04%. Levonorgestrel 0.75 mg tablet is the highest in the percentage of stock wasted due to expiry. Factors contributing to wastage due to expiration were supply and demand imbalance.
• Definition: % of stock wasted due to expiration = Unusable stock of an item during the year * 100 / (Opening stock + Quantity received of the item during the year)
• Expiration rate is provided for 25 consumables

2. Gurmu et al (2017): Inventory management performance of key essential medicines in health facilities of East Shewa Zone, Oromia Regional State, Ethiopia

• The mean stock out rate of key essential medicines was around 27.25% with average stock out duration of 35.31 days. On average around 10.43% of medicines were wasted resulting in loss of 174,366.98 Ethiopian birr (ETB) of which the value of medroxy progestrone accounted around 65.74% of the loss.

3. Masefield, Msosa, Grugel (2020): Challenges to effective governance in a low income healthcare system: a qualitative study of stakeholder perceptions in Malawi

• The government requires that all national health system drug-procurement is via the Central Medical Stores (CMS) or their approval is sought before using other sources or distributing donations (P5). When this procedure is followed the supply can be poor, sometimes drugs are available in the CMS but not received by the hospitals (P13). There were calls from the interviewees for an improved system to coordinate between the CMS and the hospital pharmacies, and better auditing of drugs at healthcare facilities (P4, P13). In reality, drugs are accessed from a variety of sources i.e. the CMS, District Health Officers, donors and disease-specific programs (P5) and CHAM all use their preferred suppliers as an alternative to the CMS (P5, P15). For example, ‘since the city assembly is autonomous, they have at times decided to purchase from preferred local suppliers’ and ‘we have prequalified suppliers each and every year then we negotiate the prices and we buy our own drugs’ (P15). Further, when the service providers receive donations of drugs (including prenatal multivitamin tablets for pregnant mothers) from ‘international well-wishers’ they are supposed to consult the government about distribution but instead, they distribute them as they see fit, according to the needs of the community (P15). Other providers refuse them as ‘the drugs received are based on donor preferences. The MoH has given District Health Officers powers to refuse drugs which are not needed because it costs more to receive drugs that will not be used’ (P19)

4. Nakyanzi et al (2009): Expiry of medicines in supply outlets in Uganda

• Study area - 19 public medicine outlets (3 non-profit wholesalers, 16 hospital stores/pharmacies), 123 private wholesale pharmacies and 173 retail pharmacies (70% of the country's pharma business).
• cross-sectional survey of 6 public and 32 private medicine outlets (semi-structures questionnaires)
• Medicines prone to expirt include those used for vertical programmes, donated medicines, and those with a slow turnover

5. Jablonski et al ()

• survey in a representative sample from areas that included 97 clinics, resulting in responses from 3,360 citizens.
• More than 80 percent think that officials can make a lot of money by stealing and reselling drugs;
  30 percent believe that officials are unlikely to get caught stealing medicines; and
  only 50 percent believe that persons found to have stolen medicines are highly likely to face consequences.

6. [Jablonski et al (2023)] (https://www.ryanjablonski.com/articles/Jablonski%20Seim%20Carvalho%20Gibson%20with%20SI%2026092023.pdf): Using Remote Tracking Technologies to Audit and Understand Medicine Theft

• estimate that 35% of medicines go missing. Most theft occurs after deliveries, presumably by public health staff. However,supply chain error is the most common cause of missing medicines.We show that patients experience higher stockouts and pay more for medicines near facilities with more theft.

7. Chikoko (2019) (News article):

• reporting on a Global Fund Office of the Inspector General (OIG) audit report released on December 9 2019
• 32 percent of TB medicine commodities, 24 percent of HIV medicine commodities and 14 percent of malaria medicine commodities could not be traced at the DHO and health facility levels at 24 out of the 25 health facilities visited
• At central level, all sampled commodities procured by the Global Fund in 2017-18 were successfully traced to electronic and manual records at the Central Medical Stores Trust [CMST] and private warehouses. However, there is limited visibility and accountability of medicines at district and health facility levels.
• It discovered malaria cases were 28 percent higher than the underlying records, treated malaria cases reported to the Global Fund were overstated by 29 percent and that suspected malaria cases tested were overstated by 13 percent.

8. Chitete (2021) (News Article)

• The CMST presentation indicated that in 2018/19 financial year, drugs worth about K4 billion were destroyed while in 2019/20 the figure dropped to around K3.5 billion and this year the expired drugs are worth about K2 billion.
• The rate of drug expiry for Malawi, according to the presentation, is above international medical stores benchmarks pegged at two percent of stock valuation per annum. In 2018/19, CMST had the benchmark at eight percent, in 2019/20 it was at 10 percent and this fiscal year it is at three percent.

9. Kaupa & Naude (2021): Barriers in the Supply Chain Management of Essential Medicines in the Public Healthcare System in Malawi

• 12 participants based on a non-probability purposive sample from suppliers of essential medicines, regulators, donors, and logistics companies in Malawi
• Barriers identified - Lack of knowledge of the pharmaceutical market; Selection (High cost of medicines and treatment; Outdated standard treatment guidelines; Narrow range of registered products); Demand forecasting and planning (Poor quality of consumption data; Lack of skilled human resources; Insufficient financial resources); Procurement(Lack of a robust procurement system; Poor specification and quantification; Lack of financial resources; Lack of human resources and skills; Weak governance and accountability mechanisms; Inadequate capacity of suppliers/manufacturers); Warehousing and Distribution (Inadequate Storage; Lack of inventory management system; Lack of human resources; Distribution infrastructure; Lack of Collaboration)

Ratios based on LMIS data

The following figure describes the method for costing equipment in the HSSP-III

• After listing out all the requirements as per the Standard Equipment List, the HSSP-III prioritised equipment from the list so that the final cost in the strategy was 8.6% of the total needed.

• This cost is spread out equally across 8 years (12.5% cost borne each year).

• The life span of equipment under <$250,000 is assumed to be 80 years because 10% is replaced every 8 years

• New equipment which is being procured as per the equipment costing tool -

• Quantity of equipment deprioritised for the HSSP-III (only 14% of the quantity was finally included). Blank and No below imply exclusion from HSSP-III:

Options for costing equipment (following Walker & Kumaranayake (2002)) -

Updated Real versus model cost

Human Resource costing
Is there a need to account for attrition when costing health workers?
Rationale: Because the TLO model does not model pre-service training but only the resource available at a given point in time, it is important to account for costs beyond salary in the Costing module.

Total cost of maintaining the size of the current health workforce = Salaries for current workforce + Cost of regular in-service training + Cost of replacing attrited workforce

Cost of replacing attrited workforce = Cost of pre-service training per student X (1 - % of health workforce recruited from abroad) X (1 + absorption rate of students into health worker posts) X Number of health workers attrited

Number of health workers attrited = Attrition rate X Size of current workforce

References:

1. Lopes et al (2017) - Rapid review - Overall, the total annual attrition rate varied between 3 and 44% while the voluntary annual attrition rate varied between 0.3 to 28%. In the SoWMy analysis, 49 countries provided some data on voluntary attrition rates of their SRMNH cadres. The average annual voluntary attrition rate was 6.8% across all cadres.
2. Berman et al (2022) - Reliable and complete data on health worker attrition was not available to calculate health worker outflow. Through comparison of MoH analyses, analysis conducted in Malawi and other low-income countries (LICs) which suggested attrition of 14% for nurses and 15% for doctors, and MoH and stakeholder consultation, a 7% attrition rate for the non-intervention scenario was selected [Lopes et al (2017)]. This includes 1% from retirement based on birthdate data in staff returns, 2% involuntary attrition, 3% voluntary attrition, and 1% study leave
3. Beniol et al (2022)
   

HRH Costing in the HSSP-III

• Pipeline model - The objective of a health worker training pipeline model is to estimate the number of health workers expected to be available in the future workforce based on current training, recruitment, and retention trends, and under various intervention scenarios. Practically, the model provides the pace of scale up required – at each program/training institution level – to meet national workforce targets.
  

• Pre-service training costs - infrastructure, faculty development, scholarship, equipment

Other resources -
1. High attrition among CHWs
2. Need for inter-sectoral action for health worker retention

Current status of cost estimation:

Validation:

Closing PR and moving relevant scripts and resource files to PR #1637