- Top of page
- AUTHORS' CONTRIBUTIONS
Vulnerable groups are populations at risk of poor physical, psychological and social health (Aday, 1993). Despite interventions that have taken place worldwide, people of the lower socioeconomic groups run at least twice the risk of premature death and severe illness as of those near the top (Wilkinson and Marmot, 2003). The poorer health of the more vulnerable people in society can be attributed to a variety of factors such as unhealthy lifestyles, social influences and general socioeconomic conditions including unequal access to health services. Differences in healthcare utilization have been long noted among demographic groups, such as minorities and immigrants (Stronks et al., 2001; Williams et al., 2010), even if there has been a slight trend of improvement in recent years (Krokstad et al., 2002; Setia et al., 2011). As the global crisis deepens, the number of people belonging in the vulnerable groups increases and social protection becomes a necessity.
Inequalities in accessibility and in other determinants in health differ within and between countries (Makenbach et al., 2008). Thus, national policies should include specific, realistic and obtainable goals. In Greece, vulnerable groups represent a growing part of the population because of various historic, political and financial events. Greece has been a destination country for immigrants for many years. The collapse of the Eastern European regimes in the 1990s made immigration to Greece a massive phenomenon. From 1988 to 2004, the number of immigrants increased fivefold, and according to the 2001 Census, they represent 7.5% of the total population in Greece. Because of its geographical position, Greece is also a natural gateway for people who want to cross over several countries in search for a better life. An unspecified number of illegal immigrants enter mainly by the Greek–Turkish boarders. Bilateral and international agreements exist, yet the asylum seeker/refugee problem remains unsolved. The largest ethnic minority in Europe is the Roma population. In Greece, formal calculations place the Roma population at about 150 000, whereas Roma associations reckon the numbers to be about 250 000–300 000 (Fundación Secretariado Gitano, 2009).
Vulnerable groups might differ in some of their characteristics, but they all have difficulties in reaching health services. Accessibility to health services is a complex and multifaceted issue. Thus, researchers explore accessibility from different points of view. One way is to compare the usage of different groups of population. For example, in a cross-sectional nationwide household survey, they examine if immigrants are using health services equally with indigenous people for equal needs (Tountas et al., 2011). They also examine the factors of non-use, such as the likelihood of a language barrier, a cultural obstacle or a prohibitively high medical cost (Dias et al., 2008). Another approach is to analyze the number of admissions in hospitals for avoidable conditions such as measles, which is a preventable immunizable infectious disease (Pappas et al., 1997). Dixon-Woods et al. (2006) introduced the term “candidacy”, which describes the ways in which people's eligibility for medical attention is jointly negotiated between individuals and health services. Finally, there are two groups of people who understand accessibility problems: patients and health professionals. In Greece, there is little evidence about the views of health professionals on the accessibility of vulnerable groups to health services (EUGATE, 2011).
The recession deepened in 2010, and the fiscal austerity measures resulted in a reduction of salaries and social protection allowances. In 2010, more than 200 000 people lost their jobs. The unemployment rate subsequently increased and reached 15.9% in March 2011 compared with the 11.7% for March 2010 and 9.3% for the same period in 2009 (Hellenic Statistical Authority, 2011). Risk of poverty, social exclusion and poorer health are the consequences of this escalated unemployment.
In our study, we tried to detect, to record and to analyze the possible barriers to access to health services for vulnerable groups, which are related to problems and deficiencies of service facilities, or are connected with their particular living conditions and their socioeconomic and cultural characteristics. Thus, the purpose of this study was twofold: to examine the difficulties faced by the professionals, when in contact with vulnerable groups, and to identify the particular needs of these people from the professionals' perspectives. This is a qualitative substudy performed as part of a larger project on the accessibility of vulnerable groups to healthcare services in Greece. The quantitative approach substudy investigated the views of members of the vulnerable groups, and it will be presented in a forthcoming paper.
- Top of page
- AUTHORS' CONTRIBUTIONS
A two-round Delphi method was undertaken to achieve the objectives of the study. The Delphi technique is widely applied in health research as a tool for solving problems in healthcare settings. It is in essence a series of sequential questionnaires or “rounds” that aim to gain the most reliable consensus of a group of experts (Linstone and Turoff, 1975). Its main characteristics are use of experts, anonymity of the responses, multiplied rounds, controlled feedback and statistical analysis (Clayton, 1997). It is chosen for this study to support (i) the ability of each participant to express views anonymously, while providing information generated by an entire group of participants, (ii) the need of heterogeneity of the participants to insure validity and (iii) the lack of geographical limitations because the questionnaires are usually completed by mail (Linstone and Turoff, 1975; Fink et al., 1984).
The success of a Delphi study rests on the combined expertise of the purposely selected participants and not on the representativeness of the sampling for statistical purposes (Powell, 2002). There are no universal rules about the minimum or maximum number of participants. The heterogeneous panel of professionals selected for our study, consisting of academics, nurses, physicians, social workers and administrative staff, working in hospitals or health centers in areas where there is a high concentration of vulnerable groups required an increased number of participants. The selection criterion for participants was a minimum of 3 years of experience working with vulnerable groups. For the academics, the sole inclusion criterion was the publication of at least one article about vulnerable groups. Diverse geographic regions (urban, suburban and rural) were represented in the panel.
Initially, 36 individuals (14 physicians, 8 nurses, 6 social workers, 5 hospital administrative officers and 3 members of the academic society) were approached in writing per mail and invited to participate. They all agreed. This number was adopted to allow dropouts at each stage and still fulfill the requirement of having a minimum of 25 experts by the end of the study to increase reliability. Both rounds were completed by 28 participants (78% responsiveness). The experts were mostly female (61%), had a mean (SD) of 13.28 (±8.48) year experience with vulnerable groups and originated from four areas: Athens, Thessaloniki (the second biggest city), Kalamata (medium-size urban city) and Messini (a rural town) (Table 1).
Table 1. Sociodemographic characteristics of the group of experts
|Gender|| || || || || || |
|Male||11||39|| || || || |
|Female||17||61|| || || || |
|Total||28||100|| || || || |
|Occupation|| || || || || || |
|Physicians||8||29|| || || || |
|Nurses||6||21|| || || || |
|Social workers||6||21|| || || || |
|Administrative staff||5||18|| || || || |
|Academics||3||11|| || || || |
|Total||28||100|| || || || |
|Location of work|| || || || || || |
|Athens||16||57|| || || || |
|Thessaloniki||6||21|| || || || |
|Kalamata||3||11|| || || || |
|Rural areas||3||11|| || || || |
|Total||28||100|| || || || |
|Age|| || ||26||60||40.42||9.56|
|Years of experience|| || ||3||30||12.46||8.83|
To initiate the study, an invitation letter and the round 1 questionnaire were sent on 15 November 2010 to the expert panel. The invitation letter explained the objective of the study and the method to be followed. The first round aimed at identifying which factors are important relative to the research question. The questionnaire consisted of three open-ended questions and a fourth part for comments and additional suggestions. All the questions in the first round of the Delphi study were pretested by experts, and feedback was sought on comprehensibility, question order effects, readability and word ordering. Modifications were made when needed. Following the suggestion of the pilot study participants, we provided a definition of “vulnerable groups” relevant to the study at the top of the questionnaire and there was an explanation of the subgroups that composed the target group of the study. These subgroups included unemployed people, poor people, immigrants, refugees and the Roma population.
The first questionnaire included three open-ended questions that are presented in Table 2. We asked the panel of experts to give three to five answers to our questions to maximize the chance of pinpointing important issues while avoiding unlimited responses that could create an excessive volume of data for analysis (Schmidt, 1997). The participants had a 3-week period in their disposal to send back the completed questionnaires via e-mail or fax.
Table 2. The first questionnaire
|Questions||Number of responses|
|1.||Name the things that make your contact with the members of vulnerable groups more difficult||110|
|2.||In your opinion, what are the problems that vulnerable groups face when dealing with the health services of the National Health System?||108|
|3.||Name at least 3 things that could improve the rendering of health services to members of vulnerable groups||112|
The results of the first round were analyzed with the purpose of creating the second round questionnaire. The 345 items, which ensued from the first round, were summarized, analyzed and sorted by content. Answers describing the same variable were grouped into one item, and there were also cases that a response that included two different issues was split, according to the topic. Finally, the 345 items were consolidated into 55 statements. The 55 statements were divided into three thematic sections: (i) difficulties of experts, (ii) difficulties of vulnerable groups and (iii) proposals for improvement. To minimize the possible ambiguity or vagueness, the questionnaire was once again pretested by experts who did not participate in the study, and feedback was sought on word ordering, question order effects and readability. Consequently, modifications were made when needed.
The second questionnaire was sent to the group of experts to determine the relative importance of each of the factors identified. A thank you letter for participating in the first round was also sent with some information about the number of responses of the first round and brief instructions for the completion of the second questionnaire. The panelists were requested to rate each of the 55 items on the basis of their significance. They were asked how essential or necessary each item was on a five-point Likert scale (1=strongly disagree, 2=disagree, 3=neither agree nor disagree, 4=agree, 5=strongly agree).
Consensus among the panelists was considered a priori the representation of >70% agreement to each data element (Hasson et al., 2000). Agreement was expressed as the percentage of respondents scoring either 4 (agree) or 5 (strongly agree). It was decided that beside the convergence, there would be a reference to other interesting outcomes such as controversies and polarization of answers.
Number of rounds
In the classic Delphi approach, four rounds are used, although in previous studies, two or three rounds are preferred (Hasson et al., 2000). Our research group decided a priori to use two rounds because most of our experts were first-line medical practitioners and nurses under time pressure, and we wanted to avoid reduction in response rates because of participant fatigue.
Verbal and written confirmation of guaranteed anonymity was given to the experts. Informed consent to participate in the Delphi study was determined by the completion of the questionnaire.
Throughout the study, data handling was limited to two researchers and the collected material was coded prior to analysis. Results from open-ended questions included in the first questionnaire were assembled into appropriate categories and synthesized for use in the second questionnaire. Quantitative analysis included the assessment of the frequency of item selection and the calculation of answer scores. All data were analyzed with IBM SPSS 18.0 software.
- Top of page
- AUTHORS' CONTRIBUTIONS
Our study indicates consensus among experts in identifying difficulties that health professionals and members of the vulnerable groups face. Bureaucracy is interwoven with public administration since the restoration of democracy in 1974 in Greece, and it is considered one of the main weaknesses of the Greek Health System (Maniou and Iakovidou, 2009). The proposal of the experts in this study to diminish the excessive bureaucratic procedures is not connected to their general consequences for productivity or to the phenomenon of corruption of public servants. It refers to the simplification of procedures to render health services accessible to vulnerable groups. In a study about Balkan immigrants in Thessaloniki, 16% of Albanians and 21.4% of Bulgarians mentioned bureaucracy as a problem in Greek hospitals (Hatziprokopiou, 2004).
Lack of health insurance, which was also named as a barrier by the panelists, can create two kinds of problems. Physicians are reluctant to order essential diagnostic tests, and uninsured people forego preventative and postpone necessary care. The role of health insurance is mentioned in many studies (Derose et al., 2009; Siddiqi et al., 2009). For example, in a study comparing the United States and Canada with regard to the role of health insurance as primary explanation of disparities in health, it was found that the existence or inexistence of health insurance coverage correlates with access to health services (Siddiqi et al., 2009).
In general, the quality of healthcare services could improve with various interventions according to the experts. Organized cooperation between primary and secondary healthcare settings as well as with non-governmental organizations seemed to be a priority because the participants expressed variously this request on the first round and the relative statement attained a high score on the second. Members of vulnerable groups mostly visit emergency departments and health centers, but comorbidity or multimorbidity is often present, requiring personalized treatment. Recruiting sufficient numbers of appropriately qualified staff is a proposal for general quality improvement, but it is significant to the vulnerable groups because they are among the main users of the public health system. Better staffing is ranked as a high priority among other studies because it is a main cause of burnout syndrome in hospitals (Iakovidou et al., 2008). Also, digital monitoring improves quality. E-health is part of a broader digital Agenda for the European Union because it can improve the quality of care and reduce medical costs (European Commission, 2010).
The group of experts agreed on the need for continuing education on topics related to vulnerable groups such as transcultural healthcare. In a study involving 18 focus groups on the issue in the UK, health professionals emphasized that they were anxious of appearing discriminatory, being culturally inappropriate or causing affront (Kai et al., 2007). As it is shown in a qualitative study in New South Wales, Australia, about the accessibility of young people, adequate training is important for providing healthcare successfully to people with particular characteristics (Kang et al., 2003).
Suggestions of the experts refer also to the initiatives that focus on the needs of vulnerable groups. General health promotion policies and programs might not have the expected results for vulnerable group members. Qualitative research in the United Kingdom suggested that there are several constraints on the development of health promotion relative to deprivation and poverty (Daykin and Naidoo, 1997). Lack of time and inadequate information might additionally negatively affect project outcomes; therefore, careful planning is required for maximizing the positive effects on the target group.
Translation of leaflets is a basic method for providing information to foreign speakers worldwide. This method has been criticized because low literacy patients are unable to comprehend the medical instructions given (Schaafsma et al., 2003). Thus, the extended use of graphs and pictograms is proposed. Evaluation on readability, comprehensibility, communicative effectiveness and pilot testing of the leaflets is important prior to applying to the public (Garner et al., 2011).
The medical interpreter issue was essential in both rounds of the Delphi study. The quality of the relationship between the health professional and the patient has the most positive effect on satisfaction and influences health outcomes (Raposo et al., 2009). A systematic review about the use of professional interpreters in health (from 1966 to 2005) reports benefits on communication, utilization, satisfaction and clinical outcomes (Karliner et al., 2007). Speaking the same language is a necessary but not sufficient condition to build a successful therapeutic relationship. Lack of awareness of the value systems of patients and the family expectations about roles and relationships might lead to serious ethical dilemmas with deleterious outcomes (Donnelly, 2000). The proposal of the experts for training people belonging to the vulnerable groups and using them as cultural mediators can be characterized as innovative.
Strengths and limitations of the study
The Delphi method has been methodologically criticized mainly for not adhering to accepted scientific principles such as psychometric validity (Mullen, 2003). On the other hand, Goodman (1987) reported that, “If a Delphi panel is broadly representative of the population and the area of knowledge, then content validity can be assumed”. Our study included experts from different educational backgrounds as well as from different job positions, originating from various geographical areas in Greece. Delphi has also been criticized for producing results that reflect the opinions of the selected participants only (Hasson et al., 2000). A test–retest reliability analysis could possibly be applied, but a group of experts is not always tolerant to this procedure. A third round for ranking the most important statements could further complete the study, but repeated rounds might lead to participant fatigue and dropouts. Parallel data collection and analysis is justified here because it not only reflects the primary purpose of seeking consensus about the particular issue but also provides insight into methodological issues of credibility and reliability (Powell, 2002). Finally, it is difficult to contrast our study with previous ones because no research was identified examining the exact same topic. However, in a European Delphi project of 16 countries about the identification of “best practice” in the delivery of healthcare to immigrants, some suggestions (intersectorial collaboration, culturally sensitive healthcare, transcultural education and access to interpretation assistance) were identical to our findings (EUGATE, 2011).