Effectiveness of pharmacy interventions in improving availability of essential medicines at the primary healthcare level

Search method

Searches for keywords ‘Drugs, Essential’ OR ‘pharmaceutical systems’ (and related terms) were combined with terms related to PHC. We then combined these findings with terms identifying research from low or middle-income countries (Figure 1). In addition to these results, we combined search parameters related to ‘Drugs, Essential’ or ‘pharmaceutical systems’ (and related terms) with terms related to specific drugs and common medical conditions in the developing setting (Figure 1). This review was conducted from February to April 2010. The reference lists of major papers identified in the primary search were searched for additional material, along with the websites of several organisations, as outlined in Figure 1. No additional attempts were made to source unpublished grey literature from individual countries.

The inclusion criteria were broad; studies were included if they examined the impact of country-level interventions on medicines availability. The study design is described where it was reported; studies were grouped as indicated below. Studies were critically appraised for their design and its appropriateness to the question, their size, completeness of reporting and the duration of follow up. Relevant points identified in this appraisal are briefly reported in Table 2 and as appropriate in the text. The aim was to review the findings of major studies, describe their context, and to discuss the applicability of their findings to other settings. Only articles whose full-text was available in English were considered.

Abstracts were initially screened for relevance to the research question (92 found); then any paper reporting data on availability without examining specific interventions (43) was excluded, though several major reviews of this nature were used for reference purposes in the introduction and discussion. Many studies consider the implications of international economic policies (such as patent protection) on medicines availability; whilst these concepts are important, this area was considered outside the scope of this review as these interventions are not country-specific, and such articles (32) were also excluded.

Medicines’ availability

Metrics used to report the availability of essential medicines vary between studies. Table 2 indicates the definitions as applicable to each study; we excluded studies assessing only affordability or general health-service utilisation. For ‘low or middle-income countries’ we used the definition of the World Bank (2010).

Classification and description of interventions

Grouping of results according to the strength of evidence for interventions

Results are grouped in Table 2 according to (i) good evidence: at least one high quality RCT or systematic review, which supports intervention and is not contradicted by evidence of similar quality; (ii) some evidence: existing evidence supports intervention but evidence is of low grade and quantity; (iii) weak/mixed evidence: evidence does not support intervention, or results conflict between studies; (iv) no published data: no systematic evaluation of intervention has been published.

Levels of evidence for individual studies

We assigned the following levels of evidence to papers included in the review:

Privatisation of distribution

Privatisation of supply chain elements has been trialled in several countries. We were unable to find an example of a developing country that had successfully privatised their entire supply system and reported improved access to essential medicines.

Four studies were found: three studies from Malaysia evaluated availability and affordability; one examined pre- and post- privatisation data, using standardised methods (Babar & Izham 2009), whilst two used only post-privatisation data (Saleh & Ibrahim 2005; Babar et al. 2007); one case-study examined contracted-out distribution systems in South Africa (Summers et al. 1998).

In Malaysia, distribution to the public sector was privatised in 1994; medicines remained free to patients in public facilities. Different studies have produced conflicting results on availability. One survey in 2005 found availability to be between 69.2% and 95% in the public sector (Saleh & Ibrahim 2005). A more recent survey of public hospitals found availability of 48 indicator medicines to be 25% overall; no historical data was presented in either study (Babar et al. 2007). In the external private sector, availability was 43% in the later study and prices were 6.6 to 16 times higher than International Reference Prices. The price of drugs to consumers rose by an average of 74% between 1994 and 2003 (Babar & Izham 2009); two of three studies concluded that public access to essential medicines had declined in Malaysia since privatisation (Babar et al. 2007; Babar & Izham 2009).

A case-study in South Africa examined a public-private partnership (PPP) for drug distribution across two provinces (Summers et al. 1998). One partnership began in 1996, achieving 93% availability by early 1998. Later, funding shortfalls meant the Government could not meet the terms of contract and in 1998, availability fell to 78%. The second province established a partnership in 1997, achieving 77% availability by March 1998, which dropped to 65% over the following three months. Historical data were not available. The authors concluded that the public-private model had been effective to some extent but that inherent problems such as poor quantification by health facilities and inadequate government cash flow had not yet been solved.

User-pays systems

Two studies specifically evaluated the effect of user-fees on drug availability: a retrospective case study from Cambodia (Akashi et al. 2004); and a patient survey study from Kenya (Mwabu et al. 1995). The case study from Cambodia examined medicines’ availability after the formalisation of user-fees at a public hospital replaced informal payments. This retrospective analysis demonstrated a subsequent increase in medicines availability of 10% and in medical materials of 30% (Akashi et al. 2004).

In Kenya, user fees were introduced in 1989; pre-introduction availability was not reported but 64% of patients were unable to get medicines due to stockouts in the first 9 months to September 1990, at which time fees were suspended. In the 7 months after, attendance at health facilities rose by 41% but there is no indication that this was due to improved availability of medicines (Mwabu et al. 1995).

Revolving drug funds

Three studies exploring this model were included. In Nigeria, availability in RDF facilities was 91%, against 37% in non-RDF facilities at the same time (Uzochukwu et al. 2002). In Ethiopia, availability in non-RDF facilities was 85%, compared to 70% in RDF facilities, based on 12 indicator items across nine health facilities (Carasso et al. 2009). In Laos PDR, availability was 85%, with no comparative data (Murakami et al. 2001). Another case study from Vietnam stated that availability had improved post-RDF without presenting data and was therefore excluded (Umenai & Narula 1999).

Drug- or disease-specific programmes

Three studies exploring the link between disease-specific programmes and medicines’ availability were included, two anti-retroviral therapy (ART) up-scaling case studies from South Africa and Malawi (Harries et al. 2007; Steyn et al. 2009); and one review of the effect of HIV/AIDS programmes (Yu et al. 2008).

The South African study examined an ART up-scaling programme which incorporated several functions into existing pharmacy structures, for example training of pharmacy staff, ordering, distribution and storage (National Department of Health 2003; Steyn et al. 2009). Its impact was assessed over 2 years: patient coverage grew tenfold, and availability of other drugs used in the treatment of HIV rose from 53% to 81%. The study concluded that the province had succeeded in incorporating the ART up-scale, strengthening the system overall (Steyn et al. 2009).

The case study from Malawi (Harries et al. 2007) investigated an alternative model where a separate supply system for ART was created. Pre-determined ART drug quantities were sent to HIV treatment facilities. Availability of ART improved to 100% after 12 months but there were no data evaluating the impact on the overall supply system. A review of the impact of HIV/AIDS programmes on health systems found that overall there was a positive impact on supply management processes where such programmes had strengthened existing mechanisms (Yu et al. 2008).

A case-study in Uganda examined an up-scaling programme to improve the availability of opioids and palliative medicines (Jagwe & Merriman 2007). The authors stated that availability had improved but provided no supporting data; this study was therefore excluded.

Supervisory visits

One high quality, randomised trial from Zimbabwe reported a supervision programme which provided teaching and guidelines on stock management and adherence to treatment guidelines (Trap et al. 2001). Pharmacy technicians were trained to supervise a sample number of health facilities in either stock management (n = 21) or adherence to Standard Treatment Guidelines (STGs) (n = 23). Supervision occurred at 0 months and 3 months. A third control group of facilities received no supervision (n = 18). Drug availability was measured on 15 indicator items selected from the EML; baseline data showed no significant differences in availability across all facilities. The results, over 12 months, showed that stock management indicators improved significantly and overall availability also improved (though less significantly) in the ‘Stock Management’ intervention group (Table 3); these improvements were achieved without punitive measures.

Staff training/continuing education

Two relevant studies were found: one study from Nepal examined a training project for PHC workers; another case study from India incorporated training into an essential medicines project (Chaudhury et al. 2005). In Nepal, primary health workers were given stock management training in provincial centres, learning basic computer skills and the principles of a new supply system (USAID: DELIVER Project 2009); in the initial phase, workers from 42 of 75 health districts in the country participated. Over 3 years, stock-outs of child health medicines fell from 23% to 9% and on family-planning commodities from 8% to 4%. The Delhi Programme, in India, incorporated training of health workers in a new Essential Medicines policy, from 1994 onwards (Chaudhury et al. 2005). A case-study showed that between 1994 and 2002 drug availability improved from <50% to an average availability of >80%; given the range of measures introduced however, it is difficult to conclude a causative link between that training and availability.

Community-directed interventions

Two studies were included. A study undertaken across seven research sites in three countries (Cameroon, Uganda, Nigeria) from 2005 to 2007 evaluated CDI for five health interventions (Special Programme for Research & Training in Tropical Diseases 2008). CDI is a broad concept; in this programme, communities held participatory meetings with health services and partners to make decisions with respect to such issues as dates and modes of interventions; commodity distribution processes, including persons responsible for distribution; and the nomination of community ‘implementers’, who were responsible for overseeing the community’s involvement, monitoring and reporting back to health services. These processes were based on earlier programmes using Community-Directed Treatment with Ivermectin (CDTi). The results showed that availability of drugs was higher and patient coverage exceeded control sites (Table 4). The review concluded that CDI was a viable option for the implementation of basic health programmes. These findings were supported by a smaller study in Tanzania (Kisinza et al. 2008).