ABCDEFHIJKLMNOPQRSTUVWXY
1
2
Improving Access to Palliative Care in LMICs- Cost-effectiveness Estimate
3
Country: Zimbabwe
4
ImpactBest caseGeom. meanWorst caseType of inputLevel of certaintyCitation(s) for input and notesComments/Uncertainties
5
Disease burden
6
Millions of days per 1000 population (rate)12.226.6712247753.642CitationhighInternational Association for Hospice and Palliative Care (IAHPC) Global Data Platform to Calculate SHS and Palliative Care Need (2015)Upper bound: maximum rate in 2015.
Lower bound: minimum rate in 2015.

Uncertainty: values are likely to change in the past 8 years; however, given that this is a rate, the difference is expected to be marginal.
7
Millions of days of SHS (total)191104.340787857CitationmediumInternational Association for Hospice and Palliative Care (IAHPC) Global Data Platform to Calculate SHS and Palliative Care Need (2015)Upper bound: maximum rate in 2015.
Lower bound: minimum rate in 2015.

Uncertainty: values are likely to change in the past 8 years. Given that this is a total, the difference is expected to slightly influence results.
8
Count of people with SHS685,560634,136586,570Citationmedium1. International Association for Hospice and Palliative Care (IAHPC) Global Data Platform to Calculate SHS and Palliative Care Need (2015)
2. https://data.worldbank.org/indicator/SP.POP.GROW?end=2022&locations=ZW&start=2015 		SHS count was reported in 2015 (1). World Bank reports an estimated 2% annual population growth rate from 2015 to 2022 (2).

Upper bound may assume that SHS increased proportional to population growth over the past 8 years. 591,000 + 591,000*0.02 %/year * 8 years = 685,560.

Lower bound may assume that the count of SHS stayed the same or declined over 8 years. 4,430 of SHS counts are attributed to malnutrition in Zimbabwe. No evidence supports a reduction in SHS counts since 2015 so it was assumed that the count remained the same as reported in 2015.
9
SHS days per person27916597CalculationhighIt's math
10
OR: No. of people in need of palliative care0​Khumalo T, Maasdorp V. The Island Hospice model of palliative care. Ecancermedicalscience. 2016 Jul 7;10:654. doi: 10.3332/ecancer.2016.654. PMID: 27563349; PMCID: PMC4970623.what measure do you think makes more sense?
11
12
Reduction of SHS through Intervention
13
14
15
Number of existing facilities providing palliative care262525CitationhighClark D, Wright M, Hunt J, Lynch T. Hospice and palliative care develop- mentinAfrica:amulti-methodreviewofservicesandexperiences.Journal of Pain and Symptom Management. 2007; 33(6):698-710.There are two ways to represent this number:

1. Number of parent organizations with PC services or,
2. Number of total branches and satellites of parent organizations with PC services.

According to APCA, Zimbabwe has 11 hospices/PC services but 25 PC services if including satellite operations separately.

I am using the latter figure as it may be more representative of palliative care availability. Taking the larger number could also mitigate errors with the figures being from a 2015 report as we may naively expect efforts to be expanding over the last 6 years (net GDP grew from $19.96 billion to $20.68 billion USD).

Best case scenario could be that PC services have improved with GDP (if we multiple 25 by the factor of net GDP growth in 8 years, we get (20.68-19.96)/19.96*25 = 0.9 more PC services).

Worst case may be that the number of services have stayed the same or decreased. Since we have not searched for evidence (to conserve time on this estimate) suggesting a reason for decreasing the amount of PC services available, we will assume the number remains the same.
16
% of people with SHS having access to guideline-aligned palliative care / trained staff50%16%5%Estimated InputlowWe can estimate an upper bound of 0.5, assuming that current organizations are covering 50% of the need already, or that a new intervention is only able to support 50% of the need during its first few years of operation. No evidence was found in a brief search on an estimated coverage of PC need by current actors.

A lower bound may consider that not all people in need are reached (a ballpark estimation of 1/10th of the upper bound being met).
17
Number of people with SHS having access to palliative care / trained staff342,780100,26629,329Calculationhigh
18
% of people with SHS in need of pain medication having access to pain medication or appropriate modes of relief (spiritual, financial aid etc)90%80%72%Estimated InputlowA naive upper bound may assume that those that have access to guideline-aligned palliative care also have access to medications (it is legal to prescribe opioids for PC in Zimbabwe so we do not need to consider legal barriers around that) at 80% of the need being covered.

A lower bound may represent barriers to accessing medications due to transport and stock shortages, or in prescribing medications due to disincentives or opiophobia of PC care workers, caregivers, or patients. I've assigned an arbitrary penalty for these considerations of 20%.
19
% of SHS cases detected by healthcare workers 80%63%50%Estimated InputlowA naive upper bound may assume that 90% of SHS need with access to guideline-aligned PC are seen.
A lower bound may assume that only 50% of SHS need with access to guideline-aligned PC are seen.
20
Proportion of people affected experiencing sufficient relief in SHS36.00%8.05%1.80%CalculationlowD. in target pop set spaces table = % of SHS cases detected by healthcare workers x % of people with SHS in need of pain medication having access to pain medication x % of people with SHS having access to guideline-aligned palliative care/ trained staff
21
SHS days reduced = people with SHS impacted by intervention x days of suffering per person reduced68,760,0008,399,2711,026,000CalculationlowCount of people with SHS x % of people affected experiencing sufficient relief in SHS x conversion of SHS to WELLBYs.

Naive conversion (not based on anything really, just ran out of time here): reduction in suffering by 10% increases WELLBY by 1 point. This is assuming that SHS is measured on a 1-10 point scale so reduction by 1 point = reduction of suffering by 10%, relatively speaking.

Upper bound: assume sufficient reduction in suffering is by 50%.
Lower bound: assume sufficient reduction in suffering is by 10%.
22
23
Estimated post intervention State
24
Benefit from training health workers to deliver palliative care​​
25
Number of health workers enrolled in a course1004520Value inputlowI made this up
26
Number of courses offered in a year444Value inputlowI made this up. Once a quarter sounds manageable for a small nonprofit
27
Total number of health workers trained to deliver PC in a year40017980Calculationhigh
28
Years of operation for the organisation555Value input​You can change this number as you see fit
29
Total number of health workers trained to deliver PC over lifetime2,000894400Calculationhigh
30
Patient caseload per health worker201613CitationmediumStaffing a Specialist Palliative Care Service, a Team-Based Approach: Expert Consensus White Paper300 patients require (3 PC physicians + 12 RNs + 2 PC social workers) x 130% to account for nonclinical responsibilities taking up 30% of dedicated time = 22 total FTEs for 300 patients --> 13.6 patients per FTE

These figures are for Canada. In Zimbabwe, due to staff shortages, there will realistically be higher patient caseloads per FTE, which is bad for health workers but good for cost-effectiveness. I model this higher figure in the "best" case scenario
31
Total patients cared for by health workers in year 18,0002,8621,040CalculationmediumCalculation logic: every additional year, as more health workers are trained to deliver PC, the total number of patients that can be cared for also increases, linearly with year
32
... in year 216,0005,7242,080Calculationmedium
33
... in year 324,0008,5873,120Calculationmedium
34
... in year 432,00011,4494,160Calculationmedium
35
... in year 540,00014,3115,200Calculationmedium
36
Total patient-years cared for by health workers over lifetime120,00042,93315,600Calculationmedium
37
SHS days per person per year27916597Calculationhigh
38
Total SHS days given palliative care by health workers33,432,5227,064,1141,515,932Calculationhigh
39
​​
40
Benefit from providing pain medication and palliative care products​​
41
Number of patients receiving pain meds and PC products in a year137,27014,1891,650Calculationhigh
42
Years of operation for the organisation555Value input​You can change this number as you see fit
43
Total patient-years receiving pain meds and PC products in a year686,34970,9468,250Calculationhigh
44
SHS days per person per year27916597Calculationhigh
45
Total SHS days given palliative care by health workers191,219,72911,673,528801,723Calculationhigh
46
​​
47
Total SHS days covered by both palliative care and pain meds & PC products
33,432,5227,064,114801,723CalculationhighLogic: a patient needs to be covered by both palliative care and pain meds & PC products to get the full treatment. So the benefit will be capped by the lower of the two values
48
Proportion of people affected experiencing sufficient relief in SHS36.00%8.05%1.80%CalculationlowD. in target pop set spaces table = % of SHS cases detected by healthcare workers x % of people with SHS in need of pain medication having access to pain medication x % of people with SHS having access to guideline-aligned palliative care/ trained staff
49
Total SHS days reduced via pain medication12,035,708568,65014,431Calculation
50
​​
51
Intervention effectiveness - direct benefit to patients12,035,708568,65014,431Value inputhigh
52
​​
53
Certainty discount - for weaker evidence100.0%70.7%50.0%totally made up​
54
Generalisability discount - country of intervention similar to country in RCT100.0%83.7%70.0%totally made up​
55
Bias discount - effect size is from potentially biased source100.0%70.7%50.0%totally made up​
56
Total discount100.00%41.83%17.50%
57
Intervention effectiveness - corrected by discount 12,035,708237,8832,525
58
59
Indirect effects and externalities
60
Positive indirect effects​
61
Household spillover %53.00%26.50%0%CitationlowHLI pain relief report53% seems optimistic, so put it as best case scenario
62
Household size432Citationmedium1. HLI pain relief report
2. https://zimbabwe.opendataforafrica.org 		Quote from report: "As discussed earlier, we add a household spillover effect of 53% (like that we calculated for psychotherapy for depression; McGuire et al., 2022). We guess a household of two additional people. We assume a smaller household size than usual because household size tends to shrink with age (see Thomas et al., 2021; Rychtaříková & Akkerman, 2003)."

Zimbabwe Data Portal (2022) indicates an average household size of 4.
63
Total benefits from household spillover 25,515,701178,3020Citation​
64
​​
67
72
Total benefits over time37,551,409416,1852,525Calculation
73
74
Costs (during pilot over course of implementation)Best caseGeom. meanWorst caseType of inputCitation for inputComments/Uncertainties
75
Cost of training health workers to deliver guideline-aligned palliative care
76
Cost to train a single health worker for a short course to learn the basics of PC$20.00$25.00$30.00CitationmediumHospice Africa - 5 day short courseI use their quoted price of 100,000 Ugandan shillings = 25.06 EUR per learner (as of 30 Oct 2023) as the geom. mean., assuming they are not making a profit
77
Revenue from training program per health worker$20.00$10.00$0.00Value input​I use Hospice Africa's course price
78
Cost to train a single health worker - after including revenue$0.00$15.00$30.00Value input​
79
Number of health workers enrolled in a course1004520Value inputlowI made this up
80
Number of courses offered in a year444Value inputlowI made this up. Once a quarter sounds manageable for a small nonprofit
81
Total number of health workers trained to deliver PC in a year40017980Calculationhigh
82
Total cost to train health workers to deliver PC in a year$0.00$2,683.28$2,400.00Calculationhigh
83
Years of operation for the organisation555Value input​You can change this number as you see fit
84
Total number of health workers trained to deliver PC over lifetime2,000894400Calculationhigh
85
Total health worker training costs over lifetime$0.00$13,416.41$12,000.00CalculationhighSeems small compared to the other cost line items below
86
This part should be in the benefits section
87
Cost of product
88
ongoingCost of PC products besides drugs (per unit) Citation​
89
ongoingCost of pain medication (per patient per year) $0.73$4.70$30.30CitationlowMSH price guide 2015Upper bound: Lancet Commission reports "$0.78 USD per capita per year" in LMICs. This seems absurd and unreasonable. We do not trust this number.
Lower bound: This is what HLI uses for their BOTEC on palliative care opioid supply. They cite "$8 per patient per 90 days", which I convert to a daily price. I simply double the unit price for the worst case scenario
90
Citationhigh
91
periodicCost of delivery of product (per unit)$0.00$2.35$30.30Estimated InputlowMade-up figures. Worst case, delivery is assumed to be same cost as pain meds
92
Total cost per patient per year$0.73$7.05$60.60Citation​Not really 365 units, but 365 days per year
93
Annual budget for PC products$100,000.00$100,000.00$100,000.00Value input​You can change this number as you see fit. It determines the number of patients receiving PC products
94
Number of patients receiving PC products in a year137,27014,1891,650Calculationhigh
95
Years of operation for the organisation555Value input​You can change this number as you see fit
96
Total PC product costs over lifetime£500,000.00£500,000.00£500,000.00Calculation​
97
98
108
Cost suggested by CE/RTP
109
onceFixed charity start-up costs$125,000.00$176,776.70$250,000.00RTM told me to use this number​Includes costs to identify region for intervention, developing and/or distributing knowledge-sharing materials and other initial activities in the ToC.
110
periodicOngoing annual charity costs/organisational overhead costs$250,000.00$353,553.39$500,000.00RTM told me to use this number​Includes cost of medications, subsidizing staff wages(?), transportation to homes for care, materials for at-home kits, among other perpetual costs.
111
Years of operation for the organisation555Value input​You can change this number as you see fit
112
Total organisational costs over lifetime$1,125,000.00$1,590,990.26$2,250,000.00Calculation​
113
114
Total costs without counterfactual adjustments$1,625,000.00$2,104,406.67$2,762,000.00Calculation
115