To evaluate the impact of opening up ownership of pharmacies in South Africa

Background Following the democratic elections in 1994 the South African private pharmaceutical services were mostly in metropolitan centred with a scattering of pharmacies in less densely populated areas. The Government introduced regulations relating to the ownership and licensing of pharmacies on the 25th of April 2003 to improve access to pharmaceutical services by removing ownership restriction to only pharmacists. Objective To assess the outcomes of the policy implementation in improving access to pharmacies. Method The register of pharmacies at the South African Pharmacy Council was analysed from 1994 to 2014. Each registration was assigned GPS coordinates using Q-GIS(V3.6) and mapped per province at a district level, following clean-up and verification of the register. New registrations were also categorised as either corporate or independent pharmacy. Population census was obtained from Statistics South Africa and used to determine the number of pharmacies per 100,000 population. Main outcome measure(s) Number of active pharmacies; Number of independent pharmacies; number of pharmacies in each district. Results The number of active pharmacies increased from 1624 at the end of 2003 to 3021 by 2014. The closure rate decreased from 137 to 86 pharmacies per year post regulations, a 37.23% reduction with a net gain of approximately 127 pharmacies per year. About 38.30% of all pre-2003 pharmacies (622 of 1624) closed by 2014. The population increase in the study period was approximately 20.66% but the overall growth of pharmacies was only 1.88 pharmacies per 100,000 population (3.55 to 5.43). Following the regulations in 2004, 23.9% of pharmacies active within the system closed between 2004 and 2014, of which, 91.7% of them were independent pharmacies. Conclusion Opening up of pharmacy ownership in South Africa increased the number of pharmacies in the country but did not result in increased access in previously less populated areas. There was still clustering of pharmacies in a well resourced areas, with a steady growth in corporate pharmacy (35%) ownership.


Background
Following the 1994 democratic elections the new Government in South Africa had the opportunity to introduce policies that ensured the availability and accessibility of cost-effective medicines to all South Africans. A National Pharmaceutical Policy Committee was established by the Government post elections in April 1994 [1], which led to the publication of the National Drug Policy [2]. The key concept related to pharmacy ownership was contained in the following statement; "Where it is deemed to be in the interests of the public, and provided that comprehensive pharmaceutical care is ensured, ownership of pharmacies by laypersons and other health care professionals will be considered [2]." It is important to reflect on the intention of the Minister in introducing the Bill to parliament in 1997 for debate. The Group Areas Act (1950) defined residential zones and confined healthcare professionals to their own ethnic communities [3]. Black pharmacists [4] who qualified in the 80's and early 90's were not allowed to own pharmacies in urban areas (defined as per the National Spatial Development Framework Draft 2018 as "Urban areas are characterised by large communities living at high residential densities, a variety of employment opportunities, and high-intensity business and commercial areas") [5], where trade was lucrative and profitable. Private pharmaceutical services were only accessible to affluent communities situated in metropolitan areas [6]. A metropolitan area is defined as a large densely populated city classified as Category A municipalities described in section 155(1) of the Constitution and Municipal Structures Act (Act 117 of 1998) [7]. The Bill sought to improve access to pharmaceutical services by removing restriction of ownership to only pharmacists. Further debate centred around the Minister's powers in determining who should own pharmacies, and ownership being determined on a need basis. Part of the motivation heard in parliament [4] was that opening up of ownership would reduce the price of medicines, promote healthy competition and create more jobs.
The Regulations Relating to the Ownership and Licensing of Pharmacies was published in Notice No. 553 of 25 April 2003 [8] where the responsibility to issue a license was moved from the South African Pharmacy Council to the National Department of Health. Unlike many low income countries where pharmacy oversight, regular inspection and law enforcement is weak [9], the South African Pharmacy Council has a well-defined and stringent process.
In most countries where deregulation was attempted, the rationale for change centred around the need for increased competition, containment of pharmaceutical expenditure, improved access to pharmaceutical care and opening of new outlets in areas of need [10]. The Österreichisches Bundesinstitut für Gesundheitswesen Austrian Health Institute (OBIG) 2006 report [10] of the European Union (EU) countries indicated that 17 of the 25 member nations operated restricted ownership of pharmacies. The study went further to do a comparative analysis of three EU countries that were regulated i.e. Austria, Finland and Spain compared to the deregulated states of Ireland, Netherlands and Norway. The study showed a strong increase in the number of pharmacies in the deregulated member states accompanied by urban clustering and fewer municipalities having access to service.
A 2015 survey conducted by the International Pharmaceutical Federation (FIP) [11] in 71 countries covering 80% of the world's population indicated that 66% of pharmacy ownership is non-exclusive to pharmacists and the balance of 34% (24 countries) were exclusive. Non pharmacist ownership ranged from state ownership to complete liberalisation. Other factors that determined ownership related to workforce capacity where the number of pharmacists may not be sufficient to cover the areas of need. Some countries have liberalisation but provide additional restrictions [11], the most frequent being restricting other authorized non-pharmacist prescribers from ownership, banning vertical integration in a supply chain, or restricting horizontal integration to prevent dominance. Strong regulated environments are built on restricted ownership to pharmacists, combined with geographic conditions [12] based on number of inhabitants per pharmacy and minimum distance from each other. This is meant to create a spread of pharmacies across geographic areas allowing for sustainability.
Challenges of restrictive ownership in Germany and Italy were brought to the European Court of Justice [11]. The court ruled that restriction with the justification of safety and quality is allowed. Two other countries, Hungary (2009) and Estonia (2015) [13], returned to regulated ownership based on professional independence of the pharmacists, lack of rural improvement, and financial unviability of the remaining pharmacies. In Africa, some countries such as Chad, Senegal, and Cameroon restrict ownership to pharmacists while Kenya and Nigeria, follow the South African model of liberal ownership. Countries with pro-competitive policies driven by competition authorities often drive deregulation [14].
In countries where ownership is exclusive to pharmacists [11] there is an understanding that community pharmacists form an extension of the healthcare system and provide an essential public service. These models exist extensively in Africa, Eastern Mediterranean, Australia and Europe. Multiple models of open ownership and restricted ownership in the United States (US) exist as in the case of South and North Dakota respectively. A 1963 state law restricting ownership to pharmacists was tested via the North Dakota Pharmacy Ownership Initiative [15] in November 2014 where a chain pharmacy group attempted to have the law repealed and lost in a public referendum. It was shown that across every key measure of pharmaceutical care including prescription prices, levels of patient care and most importantly rural access, North Dakota outperformed other states [16].
Other models of ownership which include nongovernmental organisations, charities, religious groups and humanitarian organisations [11] are found in 28% of countries surveyed in a study by the Federation of International Pharmacy (FIP). Brazil has a unique model of municipal owned community pharmacies (Farmacias Populares do Brasil) [11] dispensing medicines off their essential medicine lists and employing pharmacists. Since 2009 when Sweden liberalised pharmacy ownership the sector is dominated by chains and independents [13]. The rationale for the deregulation which included pricing, efficiency and usage of medicine were replaced by diversity, entrepreneurship and privatisation goals [13].
The aim of this research was to explore the impact of opening up of ownership on rural access and ownership type before and after the introduction of the regulation in South Africa. There is limited research in this area, especially from middle-and low-income countries. It is important to understand if policy objectives can be met, or if unintended consequences occur.

Methods
Although licenses are granted by the National Department of Health since 2004, service can only be activated with a SAPC certificate of registration. Thus, an analysis of the South African Pharmacy Council registers for the period 1994 to 2014 was conducted. The register data was cleaned, and allocation was done in terms of provinces. A verification process involving reconciling register records with Medpages [17] followed by random telephone sampling was conducted. Community pharmacies were classified and mapped as independent and corporate, and compared to the pre-2004 data. Opening and closures of pharmacies through the study period was recorded. Dates of Opening of new pharmacies, transfer of ownership, and closures are listed on the register. This was used in the year on year adjustment. Based on the registered ownership in the Council database, pharmacies were classified in terms of listed companies and non-listed. All pharmacies in the non-listed category was assumed to be Independent.
Global Positioning System (GPS) coordinates were assigned using Q-GIS (V3.6) before mapping at a district level. Population census (2001) and 2016 Community Survey to determine pharmacies per 100,000 population was obtained from Statistics South Africa (StatsSA) [18], as the previous community survey was published in 2011 and thought to be too dated to use in this study. The entire population was used as the denominator as community pharmacy serves both the private insured and the public in general. Both district and municipal information was sourced from the Municipal Demarcation Board [19]. According to the Municipal Demarcation Board [20], all major spatial restructuring of municipal/ district boundaries took place prior to the 5 December 2000 local elections. Thus, for the duration of the study (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) there was little to no restructuring. The deprivation index and quintile allocation calculation were done in 2013/14 based on the 2011 census (District Health Barometer (2016/17)) [21], and this classification was used in this study.

Results
The number of active pharmacies (Table 1)  . The more rural provinces such as the Eastern Cape (3.4%), North West (9.0%) and Limpopo (7.7%) showed a lower closure rate with the Northern Cape being most affected as 31% of new pharmacies closed within the study period.
From Table 2, it can be seen that Manguang district in the Free State showed a substantial increase in the number of pharmacies from 19 (2004) to 47 (2014). The majority are located in densely populated areas. Increases in all other districts remained low with the Xhariep district having only 5 pharmacies by 2014. Little or no improvement was seen in the sparsely populated rural settlements.
In KwaZulu-Natal most districts in Quintile 1 had marginal increases in numbers of pharmacies. The Umgungundlovu district increased by 42 pharmacies post regulation with a total of 65 located mostly within the city centre. This may be due to it being the second most populated district in the province, having both a Deprivation Index(D/I) of 2.28 and placed in Quintile 3. Despite an increase of 16 to 32 pharmacies in the Ugu district, access did not improve as new pharmacies were located where access already existed. The Ethekwini Municipality showed an improvement with most pharmacies located within or close to existing pharmacies (3.95 to 7.37 per 100, 000). KwaZulu-Natal improved marginally from 3. The Mpumalanga province showed the most improvement: Ehlanzeni (16 to 67), Nkangala (31 to 63) and Gert Sibanda (32 to 51). All three districts have large populations and are classified in Quintiles 3 and 4. There has been growth both in the city and regional service centres as well as in the populated rural areas especially in Ehlanzeni. The province started from a low base of 2.53 per 100,000 population and improved to 4.71 per 100,000.
Limpopo province grew by 13% and showed the best new pharmacy growth (10.67%) gaining approximately 13 pharmacies per year. All districts showed improvement in the number of pharmacies in both city and densely populated rural areas. Most districts have a large population base of over a million persons. The  Capricorn district improved from 1.56 to 3.98 pharmacies per 100,000. Overall, the province saw an improvement from 1.14 to 3.55 pharmacies per 100,000 population. While new pharmacies showed a comparatively low closure rate (7.7%), the combined closure of  *The overall SAIMD combines these individual domains of deprivation using equal weights * The results were produced at ward level, with the most deprived ward given a rank of 1 and the least deprived a rank of 4277 [30]. *Each district was ranked according to level of deprivation and categorised into a socio-economic quintile (SEQ) *Districts that fall into Quintile 1 (lowest quintile) are the most deprived districts. Those that fall into Quintile 5 are the least deprived (best-off) [30] Moodley and Suleman Journal of Pharmaceutical Policy and Practice Western Cape increased from 5.77 to 9.55 per 100,000 population with the City of Cape Town showing a marked improvement of 3.8 to 7.59 per 100,000 population mostly in the city and large regional centres. Also evident was the dominance of corporate pharmacy (150 new openings) compared to 144 independents. The Central Karoo and Overberg area showed little improvement with other districts improving only marginally. The province showed an average attrition rate of new pharmacies of 14%. Approximately 66.37% of new pharmacies opened in the City of Cape Town with the bulk of the balance being shared between Eden (12.7%) and the Cape Winelands (11.7%).
A summary of all the active pharmacies per province in 2014 (3021) is presented in Table 3 below. Of these, 2019 pharmacies (66.8%) opened after the regulation with Gauteng, Western Cape and KwaZulu-Natal showing increased new openings.

Discussion
The increase in the number of pharmacies post regulations from 1624 in 2004 to 3021 in 2014 (Table 3) is in keeping with the OBIG 2006 European [10] study which showed that there was an increase in the number of pharmacies in countries that had introduced liberalisation. Norway has 8500 [22] inhabitants per pharmacy with the regulated Spain (2050) and Austria (3700). Ireland, a deregulated zone has 3000 inhabitants per pharmacy. South Africa moved from 28,000 to 18,000 inhabitants per pharmacy, short of the acceptable international standards.
Most growth of new pharmacies occurred in Gauteng, KwaZulu-Natal and Western Cape. These provinces contain the major metropolitan areas; Tshwane, Ekurhuleni, City of Johannesburg; Ethekwini; and City of Cape Town. These five large districts obtained 52% of all new pharmacies. This urban clustering and lack of improvement in rural areas is in keeping with local [6,14] and international [10,11] study findings. Areas with the highest deprivation had fewer pharmacies per 100,000 population. Within the framework of current legislation South Africa must find a way to incentivise the opening of pharmacies in areas of need.
After Norway's [10] deregulation in 2001 every second municipality had no pharmacy. Urban clustering, vertical integration and chain ownership by wholesalers resulted in 4 of 5 pharmacies being owned by 1 of 3 chains. Pharmacists own only 19% of Norwegian pharmacies. The Norwegian experience led researchers to believe that deregulation leads to market dominance and minimises competition. Principle areas of practice in Europe are 78.5% in community, 8.9% in hospital and 12.6% in other areas [23]. In South Africa 68.3% of registered pharmacists practiced in the community sector in 2014 [24]. Any regulation must be carefully monitored to ensure stability and job security in this market. Deregulation in most countries [14] results in corporatisation of community pharmacy. In South Africa following deregulation 35% of new pharmacies were corporate listed. Similarly, Norway (96%), Sweden (86%), US (64%), and United Kingdom (UK) (61%) showed dominance of corporatisation post deregulation [14,25].
In Sweden the Agency for Growth Policy Analysis (Ministry of Enterprise, Energy and Communication) found that after deregulation, new pharmacies opened in urban and not rural areas, and the price of over-thecounter medicines did not decrease [26]. Lluch and Kanavos [27] highlighted the risk associated with chains and vertical integration leading to monopoly. Policies addressing these risks should be considered.
The study does have limitations. The pharmaceutical service per population ratio is only reflective for community pharmacy and excludes the public sector. The type of ownership was restricted to independent and corporate pharmacy only. The primary source document which was the Council register had inaccuracies as well as insufficient ownership data. The study did not look into quality of service provided, or operational efficiencies.
Future research should include investigating: means of improving "rural policy, rural health services and rural practice [28]"