Our searches of Medline/PubMed, EMBASE, and Global Index Medicus identified 2,481 total results (1,126 from Medline/PubMed, 1,220 from EMBASE, and 135 from Global Index Medicus). We used EndNote and Rayyan to de-duplicate the search results, resulting in 1,808 articles for abstract screening. Based on our review of article abstracts, we identified 405 abstracts as potentially meeting our criteria and were able to retrieve 310 articles for full text screening. Following full text screening, we identified 250 articles as meeting our criteria. We identified three (3) additional articles in the HAI Medicine Prices, Availability, Affordability & Price Components Database that met our criteria, the results of which were not published in articles already captured by our literature searches. This increased the total number of articles included to 253. During the data extraction process we further excluded 9 articles based on a secondary review, leaving a total of 244 articles in our final dataset. The full list of articles, including title, authors, journal title, and year published, is listed in the supplemental file. The record inclusion and exclusion decisions are described in the PRISMA flow diagram included as Fig. 1.
Types of Studies. Most articles (212/244; 87%) were descriptive studies. Only 11% (n = 28) of all the articles were intervention studies that examined the potential effect of programmatic or other interventions on medicine availability and/or affordability. The remaining four (4) articles were reviews. Of the descriptive studies, 27% (57/212) used statistical analytical techniques to identify associations between medicine availability and/or affordability measures and other variables, such as medication adherence,35 use of medicines,36 or compliance with treatment guidelines.37
Study Locations. Table 3 shows the study location(s) by country. Our results included studies from 26 out of 47 (55%) countries in the WHO Africa Region. The countries with the greatest number of studies were Tanzania (50), Uganda (49), Ethiopia (35), Kenya (33), Nigeria (29), and Ghana (21). Some articles included study sites in multiple countries. The list of study sites for each article is listed in the supplemental file.
Table 3
Study Location(s) by Country
Country
|
No. of Articles
|
Country (continued)
|
No. of Articles
|
Tanzania
|
50
|
Benin
|
5
|
Uganda
|
49
|
Senegal
|
5
|
Ethiopia
|
35
|
Burkina Faso
|
4
|
Kenya
|
33
|
Sierra Leonne
|
4
|
Nigeria
|
29
|
Madagascar
|
3
|
Ghana
|
21
|
Mali
|
3
|
South Africa
|
15
|
Zimbabwe
|
3
|
Zambia
|
12
|
Botswana
|
2
|
Malawi
|
10
|
Gambia
|
2
|
Cameroon
|
7
|
Swaziland
|
2
|
Democratic Republic of Congo
|
7
|
Burundi
|
1
|
Rwanda
|
7
|
Congo
|
1
|
Mozambique
|
6
|
Lesotho
|
1
|
Types of Medicines Studied. Table 4 shows the number of articles that analyzed medicine availability and/or affordability for different categories of medicines, based on the categorization scheme used in the 2021 WHO Model EML. The categories of medicines most commonly studied were antiprotozoal medicines (100 articles) (primarily antimalarials) and antibacterials (93 articles), followed by cardiovascular medicines (70 articles), gastrointestinal medicines (66 articles), medicines for reproductive health and perinatal care (65 articles), medicines for pain and palliative care (62 articles), and medicines for endocrine disorders (62 articles). We did not identify any articles on medicines for ectoparasitic infections, peritoneal dialysis solution, or dental preparations.
Table 4
Types of Medicines Studied
Medicine Category
|
No. of Articles
|
Medicine Category (continued)
|
No. of Articles
|
Antiprotozoal medicines
|
100
|
Diuretics
|
23
|
Antibacterials
|
93
|
Anaesthetics, Preoperative Medicines and Medical Gases
|
21
|
Cardiovascular Medicines
|
70
|
Dermatological Medicines (topical)
|
21
|
Gastrointestinal Medicines
|
66
|
Anthelminthics
|
20
|
Medicines for Reproductive Health and Perinatal Care
|
65
|
Vitamins and Minerals
|
20
|
Medicines for Pain and Palliative Care
|
62
|
Immunologicals
|
17
|
Medicines for Endocrine Disorders
|
62
|
Antifungal medicines
|
15
|
Medicines Acting on the Respiratory Tract
|
46
|
Medicines for Diseases of Joints
|
14
|
Antimigraine Medicines
|
45
|
Ear, Nose and Throat Medicines
|
7
|
Anticonvulsants/Antiepileptics
|
43
|
Antiseptics and Disinfectants
|
6
|
Ophthalmological Preparations
|
37
|
Antidotes and other Substances used in Poisonings
|
5
|
Medicines for Mental and Behavioural Disorders
|
37
|
Blood Products of Human Origin and Plasma Substitutes
|
4
|
Not specific to a particular medicine
|
35
|
Antiparkinsonism Medicines
|
3
|
Antiviral medicines
|
28
|
Diagnostic Agents
|
1
|
Antiallergics and Medicines used in Anaphylaxis
|
27
|
Muscle Relaxants (peripherally-acting) and Cholinesterase Inhibitors
|
1
|
Immunomodulators and Antineoplastics
|
26
|
Medicines for ectoparasitic infections
|
0
|
Solutions Correcting Water, Electrolyte and Acid-base Disturbances
|
25
|
Peritoneal Dialysis Solution
|
0
|
Medicines Affecting the Blood
|
24
|
Dental Preparations
|
0
|
Availability Measures. We identified 197 articles that included studies that measured medicine availability and categorized the availability measures used into six categories. The most common measure was whether a medicine was in stock on the date of a survey (i.e., cross-sectional survey) (124 articles). This included, but was not limited to, studies that used the WHO/HAI survey methodology. The second most common measure was whether stockouts occurred during a particular time period (48 articles). Often these two measures were combined. For example, Iwu, Ncobo, et al. combined the first two measures to assess the occurrence of stockouts of six tracer vaccines in Eastern Cape, South Africa.38 They assessed whether the vaccine was available on the date of a survey and in the preceding 24 months using a questionnaire, record checks, and observation. The third most common medicine availability measure was respondent self-reported availability of a particular medicine (18 articles). Aweucha, Janefrances, et al. used this approach to examine the impacts of the COVID-19 pandemic on patient access to essential medicines.39 They implemented a cross-sectional survey using electronic questionnaires across 36 states of Nigeria by asking patients whether they had “Difficulty accessing essential medicines” before and/or during the COVID-19 pandemic. Six articles used a measure of whether a medicine was on a stock list only. Five articles used a measure of the amount of stock available for a particular medicine. Kusemererwa, Alban, et al. used a stock level measure to assess medicine availability in the Uganda’s public sector.40 They measured facility stock levels and characterized stock as optimally stocked, understocked, or overstocked. An item was considered optimally stocked if the facility had two to five months of stock, understocked if it had less than two months, and overstocked if it had more than five months. To calculate the months of stock on hand, the authors divided the stock level on the day of the study by its average monthly consumption. We also identified three articles that used prescription refill data to inform their assessment of medicine availability. Table 5 summarizes the primary availability methods and measures used by the studies in our review.
Table 5
Medicine Availability Methods & Measures
Availability Methods
|
Availability Measures
|
Facility stock surveys (e.g., identified surveyor or mystery shopper)
|
Whether medicine was in stock on a certain date
|
Stock level on a certain date (e.g., stockout, low stock, medium stock, high stock)
|
Retrospective review of facility or system-level medicine stock data/records (e.g., paper bin cards, routine data systems)
|
Whether medicine was included on the facility stock list
|
Whether stockout occurred during date range (e.g., 1, 6, or 12 months)
|
Daily stock levels during date range (e.g., stockout, low stock, medium stock, high stock)
|
Prescription refill rate during date range
|
Patient or health worker self-reported availability
|
(Ask) how often certain medicines are available/unavailable based on their experience (e.g., never, rarely, sometimes, most often, always)
|
(Ask) whether they agree with a statement that describes the goal of consistent availability of medicines or the problem of unavailability of medicines (e.g., my facility has enough [insert name of medicine]; (2) I am limited by the unavailability of [insert name of medicine])
|
Affordability Measures. We identified 59 articles with studies that included measures of medicine affordability, of which the most common affordability measure compared the price of the medicine to the daily wage of the lowest paid government worker (32 articles). This is the measure used in the WHO/HAI methodology. The second most common affordability measure was patient self-reported affordability (9 articles). The methodology employed by Embrey, Vialle-Valentin, et al. included a household survey in Tanzania that asked respondents whether they “had to sell things or borrow money to pay for medicines at some time in the past” and whether the household could “usually afford to buy needed medicines.”41 Oridanigo, Salgedo, and Kebene used both a self-reported/perceived affordability measure and a standardized measure (i.e., daily wages of the lowest paid government worker) to measure medicine affordability in Ethiopia.42 Oyando, Njoroge, et al. conducted a patient survey in Kenya to assess affordability of hypertension care and asked patients if they did any of the following to cover hypertension care costs: “borrowing (having taken a loan), selling household items or assets (eg, livestock), and use of savings.”43 Five articles reported studies that used capacity to pay or similar calculation based on individual income and expenses. Khatib, McKee, et al.44 and Attaei, Khatib, et al.45 characterized medicines as affordable if the combined cost was less than 20% of household capacity-to-pay. Capacity-to-pay was calculated based on the household income remaining after basic subsistence needs have been met. A smaller number of studies measured affordability by comparing medicines prices with a population-level income standard such as per capita income (4 articles), minimum wage (4 articles), or the national poverty line (1 article). For example, Khuluza and Haefele-Abah used the statutory minimum daily wage of Malawi as the affordability threshold.46
Table 6 summarizes the primary affordability methods and measures used by the studies in our review.
Table 6
Medicine Affordability Methods & Measures
Affordability Methods
|
Affordability Measures
|
Compare medicine prices with population-level affordability standard
|
Daily wage of lowest paid government worker
|
National poverty line
|
Daily minimum wage
|
Daily per capita income
|
Calculate percentage of actual household income spent on medicines
|
Percent of actual household income spent on medicines
|
Capacity-to-pay measures (e.g., % of household income spent on medicines after covering basic subsistence needs)
|
Self-reported affordability of medicines
|
Ask whether the household could usually afford to buy needed medicines
|
Ask whether they took any steps that might indicate financial hardship to buy medicines (e.g., borrowing money, taking out a loan, selling household items or assets, or using savings intended for another purpose)
|