Results discern between descriptive and content analysis. Descriptive analysis includes: evolution in time, top 15 journals, top 10 authors. Content analysis concerns the identification of the RAs from the clusters emerging with the bibliographic coupling technique.
Evolution in time, top authors and journals
Articles on PPP in the health sector were examined in the timeframe from 1993 through 2018 (Figure 2 – blue line). From 1993 to 1999 only one item was written, but from 1998 on, attention increased and about 5 items per year were published until 2010. The first peak of 16 articles was seen in 2011, then 17 in 2015 and 23 in 2016. This may reflect changes in health systems after the global crisis when financial constraints led to large-scale reorganisation in many countries. However, the data must be considered together with the orange line showing the number of citations the articles received over time. The joint reading of the two lines gives a preliminary indication of the influence of the articles. No citations were received by the first article in 1993, so the timeframe for citations becomes 1999-2018. The two lines however provide no information on the number of citations or the lag time, which are thus shown in Figure 3.
Figure 3A shows the number of citations per year. It is noticeable that in the period 2000-2007, more than a thousand citations were received by 33 articles (Figure 2). This trend is replicated in the periods 2009-2012 and 2014-2016 (Figures 2 and 3A), but with some clear differences:
- the number of articles cited is higher; this means that each article may have lower impact than those in the period 2000-2007;
- the periods are shorter than the seven years of 2000-2007; the distribution over time of citations may mean that they are less influential in 2000-2007;
- the age of the articles is different; those published earlier might have had more chance to be read and cited. There is an issue linked to the lag time between articles published at different times, which may impact on the significance of their influence, even when the number of citations is significant.
Figure 3B attempts to resolve the issue of lag time by reporting the CPY [36, 40]. It shows the values of the CPY calculated considering the breadth of the timeframe. For this analysis, the focus moves to the two most recent periods as the influence score is higher than the first one from 2000 to 2007. Articles published in 2018 received 27 citations during the year of publication, which places those articles in third position in the overall rank of the CPY. However, a further aspect is the number of articles published, which affects the influence of each article and makes citations more difficult to interpret. Without taking this into account, the data presented in Figures 3 are biased. To solve this problem, Table 1 reports the top 10 cited articles and related CPY.
Following this rationale, it was found that the most influential articles by CPY are located in the earliest period 2000 - 2007. In fact, eight out of ten articles are from the first period and the next two articles from 2009-2012, and are moreover ranked only seventh and eighth. This kind of information is necessary to raise researcher awareness of potential problems in handling scientific archives. Moreover, the fact that these articles are at the top of the CPY ranking does not mean that they are included in the cluster analysis performed to assess the content of the articles. But the joint reading of the descriptive tables in Figures 2 and 3 is an introduction to a the subsequent more detailed investigation using bibliographic coupling for bibliometric analysis.
Table 1. Top 10 articles ranked by CPY
Authors (Year)
|
Title
|
Source title
|
Cited by
|
CPY
|
Ranking CPY
|
Buse and Walt (2000)
|
Global public-private partnerships: Part II - What are the health issues for global governance?
|
Bull World Health Organ
|
161
|
8.94
|
2
|
Widdus (2001)
|
Public-private partnerships for health
|
Bull World Health Organ
|
134
|
7.88
|
4
|
Molyneux and Zagaria (2002)
|
Lymphatic filariasis elimination Progress in global programme development
|
Ann Trop Med Parasitol
|
128
|
8
|
3
|
Hotez and Ferris (2006)
|
The antipoverty vaccines
|
Vaccine
|
119
|
9.91
|
1
|
Gupta et al. (2002)
|
Increasing transparency in partnerships for health - Introducing the Green Light Committee
|
Trop Med Int Health
|
97
|
6.06
|
5
|
Bathurst and Hentschel (2006)
|
Medicines for Malaria Venture: sustaining antimalarial drug development
|
Trends Parasitol
|
59
|
4.92
|
6
|
Widdus (2005)
|
Public-private partnerships: An overview
|
Trans R Soc Trop Med Hyg
|
52
|
4
|
9
|
Lang and Greenwood (2003)
|
The development of lapdap, an affordable new treatment for malaria
|
Lancet Infect Dis
|
47
|
3.61
|
10
|
Mavalankar et al. (2009)
|
Saving mothers and newborns through an innovative partnership with private sector obstetricians: Chiranjeevi scheme of Gujarat, India
|
Int J Gynaecol Obstet
|
41
|
4.55
|
8
|
Zhang et al. (2010)
|
Control of neglected tropical diseases needs a long-term commitment
|
BMC Med
|
39
|
4.88
|
7
|
As follow-up to Table 1, it was attempted to identify the most influential journals, and whether information from the column “source title” in the top ten articles by CPY is confirmed when considering cumulative citations of all articles published by the journals. The list below shows the top fifteen journals ranked by number of citations received by the articles published:
- Bulletin of the World Health Organization (347 citations; 4 articles)
- Tropical Medicine and International Health (199 citations; 5 articles)
- Annals of Tropical Medicine and Parasitology (128 citations; 1 article)
- Vaccine (119 citations; 1 article)
- Health Affairs (112 citations; 9 articles)
- Transactions of the Royal Society of Tropical Medicine and Hygiene (84 citations; 2 articles)
- Health Research Policy and Systems (61 citations; 4 articles)
- Trends in Parasitology (59 citations; 1 article)
- Lancet Infectious Diseases (47 citations; 1 article)
- Health Policy and Planning (46 citations; 4 articles)
- International Journal of Gynecology and Obstetrics (41 citations; 1 article)
- American Journal of Public Health (39 citations; 2 articles)
- BMC Medicine (39 citations; 1 article)
- Health Policy (38 citations; 2 articles)
- Nature Reviews Clinical Oncology (37 citations; 1 article)
The Bulletin of the World Health Organization remains in top position with 4 items and 347 citations in all, but other journals, move up or down according to the number of articles and related citations computed in the new ranking. An interesting aspect is that journals specifically focusing on policy and related economic issues enter this new ranking, although in the lower positions. This was expected, because journals focusing on Medicine/Health might be considered more multidisciplinary than those focusing on health management and economics.
Content analysis: research areas
Table 2. Clusters by bibliometric coupling of documents
Cluster
|
Authors (Citations)
|
Cluster 1 – red
|
Anderson (6) [41]; Barr (36) [2]; Buse and Walt (161) [14]; Gupta et al. (97) [42]; Hein and Kohlmorgen (17) [43]; Johnston and Finegood 27) [5]; Lo (128) [44]; Molyneux and Zagaria (128) [45]; Peters and Phillips (37) [46]; Streefland (24) [47]; Vian et al. (22) [48]; Wheeler and Berkley (36) [49].
|
Cluster 2 – green
|
Ejaz et al. (20) [50]; Holden (20) [23]; Jacobs et al. (15) [51]; La Forgia and Harding (20) [52]; McIntosh et al. (8) [28]; Palmer and Mills (20) [53]; Palmer (10) [24]; Sekhri et al. (23) [54]; Whyle and Olivier (9) [21].
|
Cluster 3 – blue
|
Abuduxike and Aljunid (9) [55]; Barker et al. (7) [56]; Bottazzi et al. (15) [57]; Bottazzi and Brown (10) [58]; Hendriks et al. (6) [59]; Hotez and Ferris (119) [60]; Mahoney (18) [61].
|
Cluster 4 – yellow
|
Lambert et al. (18) [62]; Lei et al. (13) [63]; Naqvi et al. (18) [64]; Newell et al. (14) [65]; Saw et al. (10) [66]; Tin et al. (14) [67].
|
Cluster 5 – violet
|
Kruk et al. (14) [68]; Mwisongo and Nabyonga-Orem (9) [69]; Rao et al. (12) [70]; Saxenian et al. (23) [71].
|
Cluster 6 - light blue
|
Ali et al. (22) [72]; Basit et al. (7) [73]; Nishtar (10) [15]; Nishtar et al. (17) [74].
|
Cluster 7 – orange
|
Meredith et al. (23) [75]; Zhang et al. (39) [77]; Zhou et al. (12) [76].
|
Cluster 8 – brown
|
Alemnji et al. (22) [78]; Alemnji et al. (25) [79].
|
|
|
|
The analysis of the 47 scientific articles selected through the bibliometric clustering process (Table 2 and Figure 4) identified four main RAs. The type of collaboration, the roles of the different partners and the objectives pursued by the partnership are the main elements used to assign each article to the relevant RA. Table 3 shows the distribution of 47 articles over the various RAs.
Table 3. PPP in LMICs: research areas, sections and scientific publications
Research Area
|
Section
|
Scientific publications
|
Transfer of resources
|
Tangibles resources
|
[14, 45, 48, 55, 59, 69, 70, 71]
|
Intangibles resources
|
[41, 42, 46, 73, 75]
|
Co-production of health goods and services
|
Contract-based agreement
|
[28, 51, 52, 54]
|
Non-contractual based agreement
|
[49, 57, 56, 58, 60, 72, 79]
|
Governance networks approach
|
Transnational partners
|
[15, 23, 47, 74, 43, 44; 67, 68, 76, 77]
|
Local partners
|
[24, 62, 64, 66, 78]
|
Criteria for successful partnership development
|
General framework
|
[2, 21].
|
Specific issues
|
[5, 50, 53, 61, 63, 65]
|
RA 1: Transfer of resources
The transfer of resources between actors in the same partnership is used as the criterion for including articles in RA 1. Buse and Walt have defined Global Public-Private Partnerships (GPPP) as a “collaborative relationship that transcends national boundaries” [14]. Facilitation of relationships between international and local partners is a key WHO role and Buse and Walt find WHO involvement in nine out of thirteen GPPP programs [14]. Studies by Anderson [41], Gupta et al. [42], Peters and Phillips [46], Meredith et al. [75], Hendriks et al. [59], and Basit et al. [73], demonstrate this involvement.
To identify RA 1 more clearly, a further classification was made, into those dealing with tangible and those dealing with intangible resources (Table 3). Buse and Walt make a classification of GPPPs into three categories: product-based, product development-based and systems-based [14]. Other authors focus on individual international drug transfer programs for the control of specific diseases, such as lymphatic filariasis [45], onchocerciasis [46, 75], and HIV/AIDS [70]. Equally interesting are the studies by Abuduxike and Aljunid [55] and Hendriks et al. [59] which describe technology transfer projects for the production of drugs for developing countries. Saxenian et al. report the case of the PPP created to help the poorest countries introduced new vaccines thanks to joint financing by other international partners [71]. Finally, Mwisongo and Nabyonga-Orem offer a literature review on GPPP highlighting the persistent challenges in this field [69].
The literature identifies PPPs transferring intangible resources as those working with experience, know-how and skills. Pfizer’s Global Health Fellows Program aims to promote better health by improving the service delivery capacity of local partners in poor countries [48]. Basit et al. describe several training projects for the transfer of know-how for diabetes monitoring and surveillance from western countries to Pakistan [73]. Gupta et al. describe the Green Light Committee as a multi-institutional health-base partnership with the aim of making recommendations for the control of tuberculosis [42]. Lastly, Anderson presents a PPP for the transfer of appropriate expertise for the treatment of tobacco dependence from high- to LMICs [41].
RA 2: Co-production of health goods and services
Academic articles describing PPP developed for co-producing health goods and services in LMIC health systems fall under RA 2. Here it is possible to classify two groups of articles using type of PPP (Table 3). The first group includes partnerships in which a public service is funded by a formal partnership between the government and the private sector. These PPPs involve a low number of actors, and show a clear separation of roles. The government is final payer of healthcare. The private partner is responsible for co-financing, maintaining and delivering services. In health PPP, the contractual agreement creates a level of accountability in cost management and quality that may be difficult to achieve it if the government is both the purchaser and the provider of care [28].
The second group in RA 2 contains the most articles with no clear distinction of the roles between different actors. However, unlike GPPPs, PPPs do not transfer resources, rather there is a sharing of resources to achieve a common goal. Some authors recognise such PPP as Product Development PPP [58, 60, 77]. These studies describe strong collaboration projects to develop drugs and vaccines against neglected tropical diseases [79], and provide healthcare service [72]. The advantage is that each actor contributes towards the achievement of a broader goal which a single organisation would be unable to achieve.
RA 3: Governance networks
RA 3 includes articles about governance networks. Governance Networks can be found in types of PPP in which governments run schemes based on the involvement of different stakeholders in developing strategies and making decisions [80]. The health sector is one of the most complex sectors to govern and manage and, LMICs often suffer of weak capacity to perform regulatory functions [24]. Private actors often have large resources available, as well as the power to obstruct policy interventions, and it can be the case that only through collaborative action can health issues be solved [80]. Examples of dialogue between LMIC governments and private organisations have multiplied, along with attempts to involve private partners in strategic planning for the health sector [24]. Articles included in this RA can be classified into two groups (Table 3). The first group includes Governance Networks made up of transnational and local partners, and the second group includes government and local partner networks. Hein and Kohlmorgen write that a transnational Governance Networks is needed to face global health issues [43]. They find that although WHO is the main global health organisation, the World Bank is the biggest donor and should thus have a voice with WHO on global health policies. Further examples are identified in studies addressing different contexts, such as maternal health in Africa [44, 68]; malaria and tuberculosis [47] neglected tropical diseases [77] in sub-Saharan Africa [47], non-communicable diseases in Pakistan [15, 74], infections in Eastern Asia [76] and vaccines in Myanmar [67]. These emphasise the importance of conceiving of international institutions as strategic players involved in the definition of health policies and programs.
The articles in the second group describe partnership between the public and private sectors where government is not solely responsible for delivering health policies and programs, but can rely on local organisations with complementary mandates. Alemnji et al. recommend that governments in developing countries should organise committees involving private stakeholders to develop standard manuals and policies relating to HIV diagnosis [78]. Palmer [24] describes efforts made in six LMICs in establishing consultative forums for the public and private sector to define shared health programs. In the study by Lambert et al., the Bolivian government collaborates with local private pharmacies for tuberculosis control [62]. Finally, both in Pakistan [64] and in Myanmar [66], the government and the private partners have collaborated to develop a national tuberculosis program.
RA 4: Criteria for successful partnership development
RA 4 includes eight articles about building a successful PPP. Since PPPs have become a common approach to health problems in LMICs, there is an increasing need in literature to identify the conditions which make a PPP effective. Buse and Waxman study the potential risks and benefits of PPPs, and recommended that before investing in the PPP model, governments closely investigate good partnership practices and how to leverage the private sector contribution to health development [81].
The articles in RA 4 are grouped into two sections (Table 3). The first group aims to isolate the main characteristics of successful PPP building in general. Barr identifying eight principal aspects, develops a protocol to evaluate the effectiveness of PPPs [2]. Whyle and Olivier classify partnerships into eight categories and identify characteristics and critical success factors for each of them [21].
The remaining six articles identify the criteria for a successful partnership focusing on specific issues. Johnston and Finegood find three key factors for building a successful PPP to fight obesity and non-communicable diseases [5]. The studies by Newell et al. [65] and Lei et al. [63] both focus on tuberculosis. Newell et al. identify characteristics of an effective PPP in two areas: leadership and management issues, and technical issues [65]. Lei et al. identify as crucial for PPP the areas of financial costs, governance, communications and trust [63]. The last three articles have different focuses. Mahoney proposes six determinants for PPP in health technology innovation [61]. Ejaz et al. suggest how to have successful collaboration with non-governmental organisations [50]. Finally, Palmer and Mills focus on PPP contracts examining which elements influence the nature of the contractual relationship [53].