Abstract

Traditional medicine is known to be popular in sub-Saharan Africa, where over 80% have reported its utilization. It is claimed to be easily accessible, affordable, available and acceptable, but little is known about at which stage of treatment-seeking individuals turn to traditional medicines and the resulting satisfaction once used. This is due to a paucity of quantitative demand data on how many recourses of care people take for one episode of illness, whether individuals use traditional medicines as a secondary option to orthodox medicines, and if used, how satisfied they are with results. This study presents descriptive data from fieldwork carried out on 772 households in two regions of Ghana to ascertain actions taken for self-reported episodes of acute and previously diagnosed chronic diseases. Quantitative results that show by looking merely at first recourse, use of traditional medicines is fairly low, but once second recourses are accounted for there is a doubling and tripling of incidence of traditional medicine use for acute and chronic diseases, respectively. A commonly used patient-reported outcome measurement, the EuroQol 5 Dimension (EQ5D), is used to measure satisfaction before and after traditional medicine use, to reveal significantly positive changes. The study shows that whilst individuals are highly satisfied with traditional medicine, it is more often the second recourse of treatment with a revealed preference for orthodox medicines as a first recourse. This suggests that research is needed to investigate why individuals turn to traditional medicine only as a second recourse and to clarify the insufficiencies of orthodox treatment. Policies which guide individuals to take the most efficient recourses for given symptoms, and further exploration of key reasons behind high levels of satisfaction following utilization, are encouraged.

KEY MESSAGES

  • It is important to take a holistic and dynamic look at treatment-seeking behaviour and to include traditional medicines in analyses.

  • Traditional medicines tend to be used as a second recourse and, once used, seem to be highly satisfactory according to EQ5D score. More research is needed to explore the dissatisfaction with first choices made.

  • Given the popularity of traditional medicine, policy makers should look closely at reasons for delayed treatment-seeking from professional sources.

Introduction

It is often cited that 80% of people living in sub-Saharan Africa have previously used traditional medicines (WHO 2002: 1). However, little is known about its popularity vis-à-vis ‘modern’ or orthodox medicine. Owing to a paucity of health-seeking behaviour studies which examine multiple recourses and tangible outcome measurements, researchers and policy makers are left to guess that traditional medicine and related practices are at the very least reasonably popular, especially amongst more rural populations with impeded access to formal health care, but they are unable to measure exactly how high the incidence, nor at which stage of treatment-seeking individuals use traditional medicines. Indeed, the Ghana Health Service (GHS) promotes and encourages integration of traditional medicine and its practices with its orthodox system (MOHG 2004), but does so without solid documentation of basic demand-side measures.

Compared with many other sub-Saharan African countries, Ghana's health service is relatively well developed, seeking to help the poorest through the National Health Insurance Scheme (NHIS)—based upon Britain's National Health Service—which reimburses public facilities for the cost of the most prevalent diseases such as malaria, hypertension and tuberculosis. In theory, this translates into little or no cost, at the point of consumption. In practice, affordability of medicines constitutes only one dimension of an individual’s treatment-seeking behaviour, with accessibility, acceptability and availability all recognized to be important additional contributors to utilization (Aday and Andersen 1974; Mwabu 1986; Anyinam 1987; UN Millennium Project 2005). Accessibility refers to geographic factors; for example, the time taken to reach the facility, whether transportation is available and opening hours of the facility. Acceptability describes cultural elements such as attitudes of providers towards individuals (especially towards marginalized groups such as the poor, women, individuals with particular diseases) and whether the individual believes treatment will be effective. Availability refers mostly to the strength of supply-side factors, such as whether stock-out durations are relatively short and the needed medicines are available in generic form. These four dimensions illustrate the complexity associated with treatment-seeking behaviour, especially in medically pluralistic settings in which multiple types of providers exist (Mwabu 1986), as is the case for Ghana (Twumasi 1979). Consequently, individuals with medicinal needs potentially seek care from a wide range of providers.

The overarching aim of the study was to highlight empirically the sustained importance of traditional medicine for many individuals and their satisfaction before and after use. Patterns of traditional medicine use can only be fully understood and shown to be frequent by incorporating multiple recourses in analysis, especially where self-medication and self-prescribing are by far the most popular methods of dealing with symptoms (Kroeger 1983; Brieger et al. 2004; van den Boom et al. 2004). If one looks purely at first recourses, traditional medicine use seems to be relatively low (confirming that non-traditional, or orthodox, medicine is often the first and more preferred method of medication). By expanding the analysis to look at multiple recourses and outcome measurements, it is possible to paint a more holistic picture of health-seeking behaviour, and to reveal the true popularity of traditional medicines. It is argued here that traditional medicines are well-liked and levels of satisfaction with them are high because opportunity costs are low and benefits considerable. Many also believe traditional medicines to be safer, the more ‘natural’ option, and in line with causation beliefs (Mshana et al. 2008), so ‘there’s no harm in trying’, at least in the short run. In the longer run, the effects of traditional medicine and its interaction effects with orthodox medicine still remain untested or unknown, and users rely on anecdotal evidence regarding its effectiveness and safety.

Background

Traditional medicines and policy

Traditional medicine and healers are defined by the World Health Organization (WHO) as ‘the sum total of the knowledge, skills, and practices based on the theories, beliefs, and experiences indigenous to different cultures, whether explicable or not, used in the maintenance of health as well as in the prevention, diagnosis, improvement or treatment of physical and mental illness’ (WHO 2011).

The integration of traditional medicine and its related practices with orthodox systems (public, private not-for-profit, private-for-profit) was endorsed as early as 1978, when the World Health Organization officially recognized that many relied on traditional medicine owing to its relative accessibility, affordability, acceptability and availability (WHO 2002). Combining these two broad and often conflicting systems, however, has proved a great challenge in many countries. Orthodox personnel disregard traditional practitioners as untrained personnel who often trade informally without medical training and certification. Traditional practitioners often see orthodox practitioners as impatient individuals who rush their patients in and out of their consultation rooms as soon as a diagnosis is made and a prescription given. Against this backdrop, countries have taken on the WHO recommendation to integrate the systems by developing national drugs policies and tweaking existing regulations to incorporate traditional medicines and encourage the cross-fertilization of ideas. The Ghana National Drugs Policy stipulates (MOHG 2004: 8): ‘the role of traditional medicines and traditional medicine practitioners (TMP) in the health care delivery system is recognized and the Traditional and Alternative Medicines Directorate of the GHS has been established. Efforts are directed at bringing all traditional medicine practitioners under one national organization, preparing guidelines for standards of practice and ethics, and a training manual for the profession.’ However, this has resulted in a largely top down approach to policy making, without much exploration or understanding of users’ perspectives.

Choice of medicine providers

Within Ghana's medically pluralistic system (Tsey 1997), medicines can be obtained from an array of sources. The classification of providers varies across regions and districts, but here providers are classified into the orthodox (public and private), traditional (private only) and ‘self’ (medicines obtained at home, both traditional and orthodox). The orthodox system constitutes primary, secondary and tertiary institutions—some of which specialize (e.g. maternity homes)—in addition to informal services provided by drug peddlers. Formal institutions include hospitals, health centres, community health posts, pharmacies and drug sellers, and clinics. All these institutions are required to be licensed, although enforcement is not necessarily strong or successful.

The traditional system constitutes a wide range of providers and remedies which include herbal medicines, folk knowledge, rituals and spiritual elements (Tabi et al. 2006). Practitioners may include traditional birth attendants, bonesetters, fetish priests, herbalists (non-spiritual and spiritual), spiritual diviners and circumcisioners. To date, the traditional system is largely unlicensed and very few practitioners are authorized by the state as being qualified to practise officially, but by definition those in surrounding communities have given them recognition, though often only after many years of effective practice. Much like orthodox systems, the traditional practitioner specializes in specific diseases and illnesses, each with their diagnostic and treatment structures (Twumasi 1979; Anyinam 1987; van den Boom et al. 2004). Compared with orthodox systems, however, there are spiritual elements that are little understood to those without detailed anthropologic knowledge of customs and rites, frequently local in nature (Mshana et al. 2008). Conventional scientists and medical personnel fail to fully accept this spiritual element and some also reject purely herbal remedies on account of their untested, hence unviable, powers to heal [see Sachs and Tomson (1992) for a similar case in Sri Lanka]. Added to this the high propensity to self-medicate (Kroeger 1983; van den Boom et al. 2004) and the popularity of local pharmacies and medicine sellers (van der Geest and Whyte 1988; Smith 2004; Goodman et al. 2007), and Ghanaians are faced with a complex web of possibilities when deciding what actions to take when illness strikes.

For many, traditional medicine and its practitioners are favoured for commonly cited factors related to accessibility, affordability, availability and acceptability (Anyinam 1987). Herbal remedies grow in local areas or within the compound, so they are conveniently and freely accessed, they have been tried and tested by ancestors and locals alike and are often administered by those who live in the community (cultural path dependency), and are grown abundantly on fertile lands. A study undertaken on stroke patients in Tanzania shows that causation beliefs, such as supernatural causes for illness, outweighed other factors such as cost and geographic distance for care seekers, and multiple treatment-seeking was the norm (Mshana et al. 2008). Peltzer (2000) shows that in a community in South Africa, some individuals believed that doctors were able to make the illness better but could not necessarily ‘treat the cause’. Thus, even if costs of orthodox treatment may be covered by the NHIS, the preferred choice may be to obtain additionally traditional medicines/services of traditional medicine practitioners (TMPs), which can be acquired with even fewer barriers to access and are thought to be more acceptable than using orthodox medications.

Multiple recourses of care

This vast range of healthcare providers can complicate matters when analysing care seeking behaviour. The assumption that patients call at just one point of care does not necessarily hold, especially where prescriptions specify treatment courses and initial medication does not cure existing symptoms. Thus, asking individuals simply where they last or first sought care phases out the possibility that many providers were used simultaneously or in sequence. This also applies where individuals may have tried self-medication as a first recourse, thereby delaying seeking care from a professional source.

The importance of recognizing multiple recourses is paramount for both medical and economic reasons. Individuals practising polypharmacy may be at risk of dangerous drug interaction effects, and this is applicable for modern–modern, modern–traditional and traditional–traditional regime combinations. In the case of modern–modern interactions, information on this is commonly available from professional sources and on the whole warns of potential interaction and adverse side effects. For interactions involving traditional medicines, however, there may be greater risk because most substances remain untested and unapproved. Moreover, TMPs are frequently untrained on potential toxicities. One study evaluating hypoxis and sutherlandia, both commonly used herbal remedies for human immunodeficiency virus (HIV) in Africa, cautions that active constituents of these medicines inhibit the efficacy of modern HIV drugs. Further, they increase viral resistance and the probability of drug toxicity (Mills et al. 2005). Another study outlines some potential problems for individuals with hypertension, diabetes and cancer (Winslow and Kroll 1998). Such dangers are arguably more hazardous as private actions of individuals are not always revealed to physicians. Peltzer et al. (2010) find that HIV patients in South Africa refrain from telling their doctors about traditional medicine consumption for fear of stigmatization. Economically, individuals may be making irrational choices through repeat consumption when a single source may have sufficed. This would be the case if inappropriate diagnoses are made by untrained providers, or if individuals self-medicate, as is particularly common in Ghana (van den Boom et al. 2004)

In a study carried out in Ethiopia, Flatie et al. (2009) find that although modern medicine is the preferred first choice, should medication prove ineffective individuals would turn to traditional healers. Some tried home remedies first, and strongly believed that modern medicines would not work at all for spiritual illnesses. Thus, there seems to be some evidence that both types of medicines are accepted by users, but depending on the illness, people have a preference for which to use. Kroeger (1983) terms multiple treatment-seeking as ‘healer shopping’, whereby people use multiple healers for one episode of illness, without referral from the first. This is often the case where barriers to access biomedical treatment are high (de-Graft Aikins 2005).

Satisfaction levels/outcome measures

Many studies fail to measure user satisfaction in a tangible manner. This is largely because most efforts to document effectiveness have been focused on scientific results to show herbal properties, rather than the user level. Of those which do ask about satisfaction, they seek to obtain an answer to the question, ‘how satisfied were you with the service received’, or ‘how satisfied were you overall’. Although this assesses an outcome, it would be better complemented with measures incorporating a before-and-after effect to measure the change in satisfaction resulting from treatment and to avoid assuming that all individuals began at the same level or extent of health problem.

Stekelenburg et al.'s study in Zambia asked respondents about their opinions and perceptions of healer service satisfaction, although detail is omitted about the precise methodology (Stekelenburg et al. 2005: 77). The results show that 74% of respondents using traditional healers were satisfied with their treatment, but 86% of those who were not satisfied would opt for hospital treatment in the future.

To this end, the Euroqol 5 dimensions measure, EQ5D (Szende et al. 2007), is widely acknowledged as a standardized outcome measure of orthodox medicines and health service use, asking individuals to self rate on a scale five aspects of health and overall quality of life and health before and after treatment/traditional medicine use. This instrument is a simple but concise way to rate user satisfaction and health-related quality of life when combined with a corresponding question on overall satisfaction.

Methods

To better understand individual health-seeking behaviour and its outcomes, a study consisting of 772 questionnaires at the household level (4713 individuals) in two regions of Ghana was carried out in late 2010. The regions, Greater Accra (GA) and Upper West (UW), were chosen purposively for their contrasting characteristics, and where applicable, results for the two regions are presented separately. Greater Accra would be considered urban, cosmopolitan and affluent in nature, whilst Upper West is situated in a remote corner of Ghana, with many parts beyond the reach of government assistance.

A standardized methodology for household selection from WHO was used as a foundation, whereby firstly public reference facilities within a region are randomly selected, and then clusters of households in specified radii are identified. For the study, the two regions were divided into four districts including the two district capitals. In each district the regional hospital (located in the district capital) was purposely chosen and three additional reference facilities were selected, resulting in 16 reference points. Care was taken to ensure that no two selected facilities were located in the same neighbourhood. This was made possible by listing all facilities by district in Excel and applying the random number generator until facilities were spread over a mix of areas.

From each reference facility, household clusters located within three specified radii were targeted for interviews. The first radius represented a geographic location of 0.5–5 km away from the reference facility. The second radius encircled households located 5–10 km away, whilst the third radius represented locations of over 10 km from the facility. All distances were measured using a Global Positioning System (GPS) and calculated in a straight line, as the crow flies. Outside the reference facility, a bottle was spun to determine the direction of travel. Researchers walked or were driven in this direction until they were within the specified radius, from where a random household was picked by looking amongst the most common type of housing in the area. This representative household was then labelled as the starting point, and a general rule of thumb was applied to select subsequent households. Households were: at least five apart; similar to the representative household; private, not public buildings; and in separate compounds to obtain information as diverse as possible. If a household was empty or unavailable for interview at the time, researchers were instructed to return later. These rules were followed until the required number of households—approximately 16 per radius—were interviewed (resulting in roughly 48 households per reference facility).1

Internally, all members who had been residing in the household in the past consecutive month (unless newborn) were included and a suitable respondent was picked to answer questions on behalf of the household.2 This person fulfilled at least three of the following criteria: main health care decision-maker most knowledgeable about health of household members; most knowledgeable about health expenditures of the household; most knowledgeable about health utilization by household members; designated care-giver for sick household members.

Prior to the actual study a pilot was run, and ethical approval was sought from institutional affiliations in both England (London School of Economics, London) and Ghana (Kwame Nkrumah University of Science and Technology, Kumasi). All interviews were carried out by local research assistants, with prior training both in the classroom and in the field. Translations and back translations using local dialects (Ga and Twi in Greater Accra, Waale and Dagaare in Upper West) and English were made to ensure accuracy and consistency, but ultimately answers were recorded in English. Coded data were then entered into Excel, checked for inconsistencies and then transferred to STATA for further cleaning and analysis. In addition to analysis of raw data which looks purely at individual behaviour within the sample population, weighted data are also presented for the number of recourses to treatment sought and changes in dimension scores (details of this can be found in the footnote of Appendix Table 1) in order to present a regional perspective.

Questionnaire

The questionnaire was formed of modules of the following parts: basic socio-economic data of every member of the household; pregnancy module detailing all those who had used a traditional birth attendant for their last pregnancy issue/child birth; acute module for those with recent acute episodes; chronic module for those who had ever been diagnosed with a chronic illness; and finally, a module gauging attitudes and beliefs about traditional medicine followed by the household's economic and social status. For the purposes of this paper, only the acute and chronic modules were chosen to be pooled for detailed analysis of traditional medicines/practitioner use. Both modules are analysed separately in the results section below.

Users of acute TMP services were asked to rate their satisfaction, choosing one from the following: very dissatisfied; dissatisfied; neutral; satisfied; and very satisfied.

Individuals were then asked two broad questions plus five more detailed questions related to everyday life, based on the EQ5D (Kind et al. 1998). The first of the two broad questions asks how the individual rates life prior to and after traditional medicine/TMP use on a scale of 0 to 10, where 0 indicates absolute dissatisfaction and 10 absolute satisfaction, whilst the second asks about overall health prior to and after use. The difference in the before and after values is then calculated, with positive and negative scores indicating increases and decreases in well being, respectively. The EQ5D is a commonly used predictor of subjective well being and is used for cost-effectiveness analysis (Dolan et al. forthcoming). It is designed to cover as many health domains as efficiently as possible in a time/money pressured environment. These dimensions enquire about an individual’s mobility; ability to self care; ability to carry out everyday chores and activities; level of pain; and finally, anxiety and stress. In these five dimensions, individuals rate whether they have no problems, some problems or big problems (scaled 1, 2, and 3, respectively). Again, a before-and-after score is calculated, with positive values indicating positive changes, negative values negative changes and 0 denoting no change.

Results

Background data

Data from 387 households in Greater Accra and 385 households in Upper West were collated, with approximately the same number of households per radius, totalling 772 households consisting of 4713 individuals. Thus the mean number of individuals in a household was 6.1. This is in line with the figure in a similar recent report on the entirety of Ghana, which found average household size to be 6 (WHO and MOHG 2009: 13).3 Households were larger and younger in the Upper West, where there were more children (average age of all individuals 24.8 vs 26.2 years in Greater Accra). Household food and total expenditures tended to be higher in Greater Accra, where many were self-employed/had their own business, as opposed to the Upper West where the majority were farmers and subsistence farming was common. Over half the sample were female in both regions, with 56.4% in Greater Accra and 51.1% in Upper West. The main religion of those living in Greater Accra was Christianity, whereas those in Upper West mainly practised Islam. Ethnicity was roughly divided into Ga/Dangbe dialect speakers in Greater Accra and Waale/Dagaare speakers in Upper West. Table 1 outlines these sample characteristics.

Table 1

Characteristics of surveyed households

 Greater Accra Upper West 
Number of households sampled by radius:   
    1 124 124 
    2 145 137 
    3 118 124 
    Sample size 1976 2737 
    Number of households 387 385 
    Average household size 1976/387 = 5.1 2737/386 = 7.1 
    Average age 26.2 24.8 
    % females 56.4 51.1 
    % children (aged <18 years) 37.8 45.9 
    Average food expenditure in past week (cedi)a 69.9 28.2 
    Average total expenditure in past month (cedi)b 391.8 171.8 
Religion (%)   
    Christian 86.5 53.3 
    Muslim 6.6 40.1 
    Traditionalist/African or indigenous religion 0.4 3.1 
    Spiritual/African independent church 3.3 0.0 
    Combination 0.0 0.0 
    No religion/atheist 3.2 3.5 
    Other 0.0 0.0 
Ethnicity (%)   
    Akan 14.3 0.0 
    Ga/Dangbe 62.9 0.1 
    Ewe 15.3 0.0 
    Hausa 3.9 0.1 
    Dagbani 0.3 0.0 
    Nzema 0.1 0.0 
    Waale 0.0 33.0 
    Sissale 0.0 0.2 
    Dagaare 0.0 65.9 
    Other 3.3 0.7 
 Greater Accra Upper West 
Number of households sampled by radius:   
    1 124 124 
    2 145 137 
    3 118 124 
    Sample size 1976 2737 
    Number of households 387 385 
    Average household size 1976/387 = 5.1 2737/386 = 7.1 
    Average age 26.2 24.8 
    % females 56.4 51.1 
    % children (aged <18 years) 37.8 45.9 
    Average food expenditure in past week (cedi)a 69.9 28.2 
    Average total expenditure in past month (cedi)b 391.8 171.8 
Religion (%)   
    Christian 86.5 53.3 
    Muslim 6.6 40.1 
    Traditionalist/African or indigenous religion 0.4 3.1 
    Spiritual/African independent church 3.3 0.0 
    Combination 0.0 0.0 
    No religion/atheist 3.2 3.5 
    Other 0.0 0.0 
Ethnicity (%)   
    Akan 14.3 0.0 
    Ga/Dangbe 62.9 0.1 
    Ewe 15.3 0.0 
    Hausa 3.9 0.1 
    Dagbani 0.3 0.0 
    Nzema 0.1 0.0 
    Waale 0.0 33.0 
    Sissale 0.0 0.2 
    Dagaare 0.0 65.9 
    Other 3.3 0.7 

Source: author's own.

Notes: Sample size as % of region: 0.05 (1976/4,281,137) for Greater Accra, 0.41 (2737/671,043) for Upper West.

aMean value, week before survey (1st week September in GA, 1st week October in UW).

bMean value including food, month before survey (August in GA, September in UW).

Morbidity

Respondents were asked to recall any acute or chronic illnesses. The definitions and wording of both were taken from WHO approved/standardized questionnaires which have recently been used in Ghana (WHO and MOHG 2009). Acute illnesses are defined as ‘a condition that appears suddenly; the person did not have it immediately before becoming ill‘, and explained to be any disease or illness that starts suddenly and is short lived, which may dissipate with time without intervention. Classifications were double-checked using some specific symptoms for all episodes experienced by household members in the previous 2 weeks. Acute illnesses were reported by 9.8% (460/4713) of individuals. The most common acute complaint was ‘fever, headache and hot body’ (334/460, or 72.6%), followed by ‘pains and aches’ (203/460, or 44.1%) and thirdly, ‘diarrhoea, vomiting, nausea, could not eat’ (163/460, or 35.4%). In many cases this was suspected to be symptoms related to malaria.

Similarly, respondents were asked, ‘Has anyone in this household ever been told by a doctor or other health care provider that they have a chronic disease? A chronic disease is an illness that will not go away or takes a long time to go away, even when treated’. This question is designed to ascertain diagnoses made by both ‘professional’ and ‘non-professional’ providers, thus includes TMPs, and in the majority of cases individuals were able to name the illness immediately. Chronic diseases were reported by 6.4% of the sample (301 cases). Of the named chronic diseases, hypertension or high blood pressure topped the list with 77 episodes (25.6%), followed by arthritis or chronic body pain (61/301, or 20.3%) and ulcers or chronic stomach pain (46/301, or 15.3%). A combination of diseases, termed ‘other’, was also prominent, making up 49 disease episodes (16.3%). This included sickle cell, tumour/fibroid, chronic menstrual pains, hernia, ‘spiritual illnesses’ and skin conditions.

Twenty individuals were reported to have experienced an acute illness in the past 2 weeks in addition to coping with a chronic disease, whilst 3972 (84.3%) had neither type of illness.

Care-seeking behaviour

First recourse: acute

Grouping all acute episodes together, there does not seem to be an association between perceived seriousness of illness and whether care was sought, with or without medical personnel. Forty-one of 460 cases (8.9%) sought no outside care (Figure 1). Fifteen (3.3%) sought no care whatsoever, citing reasons such as ‘too expensive’, ‘care not wanted’, ‘does not trust providers of care’, or ‘there was no one at home to take me to seek care’. The remaining 40 (8.7%) who did not seek care outside the home chose to self-medicate with medicines obtained within the household (23 with orthodox medicines, 17 with traditional medicines). The majority of the sample (88.0%, 405 individuals) turned to health providers (both ‘informal’ and ‘formal’) and of these, 367 (79.8%) went to an orthodox medicines provider (termed here as ‘public’ or ‘private’ providers specializing in medicines and/or health care provision), whilst 28 used traditional medicine providers. Ten bought medicines from other unofficial medicines vendors such as a provisions shop, or a neighbour. In sum, orthodox medicines was utilized by 400 individuals (representing 90% of those who took medicines) whereas traditional medicines were used by 45 individuals (representing 10% of those who took medicines) as a first recourse.

Figure 1

First recourses taken, acute illness

Figure 1

First recourses taken, acute illness

First recourse: chronic

Turning to chronic diseases, only 5 of 301 (1.7%) individuals had not obtained medicines from outside the home in the past month to treat their diseases (Figure 2). Reasons cited for not taking medicines as recommended included: ‘symptoms have got better’; ‘care has been available at home’ (stock of medicines available at home and alternatives cost too much/are too far/are a time burden); and that ‘there was nobody at home to accompany the sick person to a provider/to obtain medicines’. The remaining 296 obtained medicines as follows: 251 (83.4%) from orthodox providers; 38 (12.6%) from traditional providers; 6 (2%) found medicines at home (‘self-medicate’); and one sourced medicine elsewhere. Figures for obtaining medicines at home (without outside care) were much lower for chronic than for acute illnesses, but those who did self-medicate, for the most part, utilized traditional medicine (five out of six cases).

Figure 2

First recourses taken, chronic illness

Figure 2

First recourses taken, chronic illness

Multiple recourses of care

Incidence of traditional medicine/TMP use changes dramatically when incorporating multiple recourses of care.

Table 2 shows a breakdown of the number of recourses sought by region and pooled. Data are unweighted4 and percentages (of row total) are given for ease of comparison. In 445/460 (96.7%) acute cases, individuals chose at least one source of care, 121 (26.3%) chose at least two providers, 13 (2.8%) turned to three different sources and 2 (0.4%) received a fourth line of care. Individuals were more likely to seek care from multiple providers when faced with chronic illnesses, with 41 (13.6%) people seeking at least three sources. Splitting the sample by region, more individuals in Greater Accra than Upper West used at least one source of care (as opposed to at least two, three or four sources) for both acute and chronic illnesses. Consequently, a larger proportion of those with needs in Upper West used two or more recourses vs those with needs in Greater Accra. Figures excluding self-medication are also given in columns headed ‘excl SM’.

Table 2

Total number of recourses to care sought by acute and chronic illness and region, unweighted

 No. with needs Number of recourses sought
 
1+ incl. SM 1+ excl. SM 2+ incl. SM 2+ excl. SM 3+ incl. SM 3+ excl. SM 4+ incl. SM 4+ excl. SM 
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) 
Acute           
    GA 185 2 (1.1) 183 (98.9) 160 (86.5) 41 (22.2) 32 (17.3) 5 (2.7) 5 (2.7) 0 (0.0) 0 (0.0) 
    UW 275 13 (4.7) 262 (95.3) 245 (89.1) 80 (29.1) 61 (22.2) 8 (2.9) 8 (2.9) 2 (0.7) 2 (0.7) 
    Total 460 15 (3.3) 445 (96.7) 405 (88.0) 121 (26.3) 93 (20.2) 13 (2.8) 13 (2.8) 2 (0.4) 2 (0.4) 
Chronic           
    GA 178 1 (0.6) 177 (99.4) 172 (96.6) 63 (35.4) 41 (23.0) 8 (4.5) 8 (4.5) 0 (0.0) 0 (0.0) 
    UW 123 4 (3.3) 119 (96.7) 118 (95.9) 76 (61.8) 80 (65.0) 33 (26.8) 32 (26.0) 5 (4.1) 5 (4.1) 
    Total 301 5 (1.7) 296 (98.3) 290 (96.3) 139 (46.2) 121 (40.2) 41 (13.6) 40 (13.3) 5 (1.7) 5 (1.7) 
 No. with needs Number of recourses sought
 
1+ incl. SM 1+ excl. SM 2+ incl. SM 2+ excl. SM 3+ incl. SM 3+ excl. SM 4+ incl. SM 4+ excl. SM 
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) 
Acute           
    GA 185 2 (1.1) 183 (98.9) 160 (86.5) 41 (22.2) 32 (17.3) 5 (2.7) 5 (2.7) 0 (0.0) 0 (0.0) 
    UW 275 13 (4.7) 262 (95.3) 245 (89.1) 80 (29.1) 61 (22.2) 8 (2.9) 8 (2.9) 2 (0.7) 2 (0.7) 
    Total 460 15 (3.3) 445 (96.7) 405 (88.0) 121 (26.3) 93 (20.2) 13 (2.8) 13 (2.8) 2 (0.4) 2 (0.4) 
Chronic           
    GA 178 1 (0.6) 177 (99.4) 172 (96.6) 63 (35.4) 41 (23.0) 8 (4.5) 8 (4.5) 0 (0.0) 0 (0.0) 
    UW 123 4 (3.3) 119 (96.7) 118 (95.9) 76 (61.8) 80 (65.0) 33 (26.8) 32 (26.0) 5 (4.1) 5 (4.1) 
    Total 301 5 (1.7) 296 (98.3) 290 (96.3) 139 (46.2) 121 (40.2) 41 (13.6) 40 (13.3) 5 (1.7) 5 (1.7) 

Source: author's own.

Note: SM indicates self-medication, percentages given as proportion of row total. GA = Greater Accra. UW = Upper West.

Individuals were questioned on whether any of the sources sought were the provider located nearest to their household. In over half of the cases, the provider to which they turned was not, offering evidence of multiple by-passing behaviour (even many second, third and fourth choices were not the nearest source of care).

When allowing for multiple recourses, 103 out of 460 cases, representing 22.4% of all acute cases, used traditional medicine and related practices. The numbers for chronic diseases were even greater, with 129/301 cases, representing 42.9%, experimenting with traditional medicines. These figures compare with those previously given for acute and chronic illness (45/460 and 43/301). In sum, including multiple recourses in the analysis more than doubles and triples the incidence of traditional medicine/TMP use in acute and chronic cases, respectively. This also suggests that traditional medicine is not often the first source to which individuals turn, leaving this option as a second, third or fourth choice.

Satisfaction levels/outcome measures

Acute

In response to the question asking users of traditional medicine/TMP services to rate their satisfaction, 76.7% were satisfied or very satisfied with the effects of treatment. Fifteen per cent felt neutral, whilst the remainder was dissatisfied. Those pleased with the results claimed that herbal treatments cured them of symptoms previously experienced, with many being able to return to their daily activities very soon after use of medicines. Individuals who were dissatisfied remarked that symptoms were still present, that they still felt the pains and that although at times the pains were relieved, sickness sometimes recurs.

When before-and-after satisfaction scores were asked for on a numeric scale, the median value for differences in health scores was +5 and the mode +6. Differences in life scores were calculated as: median +5, mode +5. Results for the five dimensions are given in Table 3, and show only three cases where conditions worsened, with the remaining 100 cases all reporting either no change or at least a small positive change.

Table 3

Changes in dimension, by type of illness, unweighted

 Unweighted changes in dimension
 
 Acute +  Chronic Acute Chronic 
Mobility       
    −2 0.9 0.0 1.5 
    −1 0.4 0.0 0.8 
    0 88 38.1 43 40.2 48 36.6 
    1 100 43.3 44 41.1 60 45.8 
    2 40 17.3 20 18.7 20 15.3 
    Total 231 100 107 100 131 100 
Self care       
    −2 0.9 0.0 1.5 
    −1 0.4 0.0 0.8 
    0 89 38.5 43 40.2 49 37.4 
    1 101 43.7 46 43.0 57 43.5 
    2 38 16.5 18 16.8 22 16.8 
    Total 231 100 107 100 131 100 
Activities       
    −2 0.9 0.9 0.8 
    −1 0.9 0.0 1.5 
    0 72 31.2 35 32.7 39 29.8 
    1 92 39.8 41 38.3 55 42.0 
    2 63 27.3 30 28.0 34 26.0 
    Total 231 100 107 100 131 100 
Pain       
    −2 0.9 0.9 0.8 
    −1 0.9 0.0 1.5 
    0 43 18.6 16 15.0 27 20.6 
    1 101 43.7 50 46.7 55 42.0 
    2 83 35.9 40 37.4 46 35.1 
    Total 231 100 107 100 131 100 
Anxiety       
    −2 0.9 0.0 1.5 
    −1 1.3 0.9 1.5 
    0 76 33.0 37 34.6 41 31.5 
    1 114 49.6 51 47.7 67 51.5 
    2 35 15.2 18 16.8 18 13.8 
    Total 230 100 107 100 130 100 
 Unweighted changes in dimension
 
 Acute +  Chronic Acute Chronic 
Mobility       
    −2 0.9 0.0 1.5 
    −1 0.4 0.0 0.8 
    0 88 38.1 43 40.2 48 36.6 
    1 100 43.3 44 41.1 60 45.8 
    2 40 17.3 20 18.7 20 15.3 
    Total 231 100 107 100 131 100 
Self care       
    −2 0.9 0.0 1.5 
    −1 0.4 0.0 0.8 
    0 89 38.5 43 40.2 49 37.4 
    1 101 43.7 46 43.0 57 43.5 
    2 38 16.5 18 16.8 22 16.8 
    Total 231 100 107 100 131 100 
Activities       
    −2 0.9 0.9 0.8 
    −1 0.9 0.0 1.5 
    0 72 31.2 35 32.7 39 29.8 
    1 92 39.8 41 38.3 55 42.0 
    2 63 27.3 30 28.0 34 26.0 
    Total 231 100 107 100 131 100 
Pain       
    −2 0.9 0.9 0.8 
    −1 0.9 0.0 1.5 
    0 43 18.6 16 15.0 27 20.6 
    1 101 43.7 50 46.7 55 42.0 
    2 83 35.9 40 37.4 46 35.1 
    Total 231 100 107 100 131 100 
Anxiety       
    −2 0.9 0.0 1.5 
    −1 1.3 0.9 1.5 
    0 76 33.0 37 34.6 41 31.5 
    1 114 49.6 51 47.7 67 51.5 
    2 35 15.2 18 16.8 18 13.8 
    Total 230 100 107 100 130 100 

Acute+chronic column do not sum to separate acute and chronic columns because some individuals had both types of illness.

Chronic

61.2% of respondents were satisfied or very satisfied following traditional medicine/TMP use for chronic illness. Another 24% were neutral about its effects, whilst 19 (14.7%) were dissatisfied or very dissatisfied. It was often the case that users of traditional medicine/TMP were on multiple sources of medication and had tried traditional medicine either as a last resort, or ‘just in case’ it would produce good results. Most users felt pains reduced, their movement was much improved and daily chores were no longer causing them problems. Those who were dissatisfied were disappointed with recurring symptoms, and general inefficacy of the herbal product. However, believers of traditional medicine were strong advocates for its effectiveness and ‘reliability’, claiming that it had better results than prescribed orthodox medicines.

The median value for differences in health scores was +5, and the mode +5. Differences in life scores were: median +4, mode +5. A handful of cases reported negative outcomes, but as with acute cases, most respondents were pleased with the effects of traditional medicine.

In sum, in both acute and chronic categories, very few experienced a worsening in any of the five dimensions (indicated by a negative change, see Table 35). There were more incidents of no changes, but the majority indicated small or larger improvements in condition following traditional treatment. Percentage scores of the totals are given in the column immediately to the right of count figures.

Limitations

This study relied on individuals’ full recall and revelation of treatment-seeking behaviour. By using self-reported health for acute illnesses, the study may have under-reported incidences if an individual had ‘normalized’ an illness. For example, because malaria is so common, they may not see it necessarily as something to report in particular, but a part of ‘everyday life’. However, in the context of a household survey, subjective reporting of illness does not seem an unreasonable way to elicit answers, if only because it is impracticable to ask for written confirmation or diagnoses, especially when self-medication is high (thus no professionals were called upon) and many people are illiterate and rely solely on verbal communication. When asking for satisfaction scores, the numbers line and the meaning of corresponding scores had to be explained in full detail. Thus, trying to objectify a somewhat subjective matter, at times, was difficult where respondents were not fully numerate. In this sense, the study would have benefited from some objective measurement [such as that introduced by the Picker Institute (Sizmur and Redding 2009)] as an additional proxy for patient satisfaction. In the two regions surveyed, different research assistants were used which may have also obscured the results. However, the consistent scores across all EQ5D dimensions suggest that the overall picture from those using traditional medicine was similar, showing high levels of user outcomes.

Overall, the study sought to be an exploratory study based on the sample population using pre-approved WHO methodology. The focus was therefore to investigate revealed individual preferences regarding health-seeking behaviour rather than undertake a region-wide analysis. In order to take the latter into account, however, weights were used and results presented as appropriate. It should be borne in mind that these weights rely on accurate data and are rudimentary in nature. As such, policy makers interested in the potential larger-scale effects of multiple recourse-seeking and changes in satisfaction should evaluate these numbers with caution. In so much as this study serves as initial evidence to inform general policy-making, it is hoped that this might encourage more research in this area.

Discussion and conclusion

This study has sought to show that traditional medicine use and related practices are still very popular amongst Ghanaians according to user incidence and satisfaction levels. There is a shortage of demand-side (household level) data in Ghana that looks at traditional medicine use, as the focus has to date been entirely on the supply-side perspective (e.g. how to integrate traditional healers into orthodox systems). As such this study is a small step towards understanding individual behavioural reactions to illness and disease to fill this gap. By taking into account up to four recourses, the study considers health-seeking behaviour as a chain of actions and reactions rather than a one shot decision. We found that in many cases people feel it necessary to utilize two sources rather than just one. This is in line with Mshana et al. (2008), who show that individuals often seek care from many sources. Multiple treatment-seeking is particularly problematic in Upper West, pointing to deficiencies in the care system in this area of the country. Similarly, individuals in more remote areas (far away from health facilities) are perhaps less likely to seek professional care, and therefore more likely to utilize traditional medicine. Such areas of Ghana are both frequently targeted and neglected; in this case, policies that address unfulfilled needs or wants following the first recourse, including those in the orthodox system, would be highly beneficial. In line with this, policies that guide individuals to taking the most efficient and best recourse for given symptoms, such as basic education and information sessions, would also reduce the number of providers sought.

The finding that most people turn to traditional medicine only as a second recourse suggests that looking only at first choices is insufficient, as the use of traditional medicine would remain hidden. By revealing second recourses, the incidence of traditional medicine/TMP in acute and chronic cases is more than doubled and tripled, respectively. This suggests that using traditional medicine is a second-best choice to orthodox medicines, but similarly it is used because of low opportunity costs and individuals feeling that they have nothing to lose by trying it. Most finish their treatment seeking within two recourses, with very few looking for third or fourth lines of treatment. Delaying treatment-seeking from professionals is thought to be common for those who have firstly tried self-treatment, but is also the result of a lack of basic knowledge and education in handling symptoms (Meyer-Weitz et al. 2000) and can result in longer-run health problems. One recommendation for policy, then, would be to consider and account for the fact that individuals are using multiple sources of care rather than merely assuming one source is used, as is commonly done in household-level surveys. Following this, policy makers would be able to question the reasons for using multiple sources—and in the revealed order—to better understand shortcomings within the system from the user's point of view.

The results suggest that individuals are engaging in polypharmacy by consuming multiple drug regimens. This can potentially lead to dangerous side-effects and inhibit the effectiveness of some drugs (Winslow and Kroll 1998; Mills et al. 2005). Policy might target encouraging patients and physicians to better engage and communicate so as to lessen asymmetry of information. For example, Howell et al. (2006) show how over 70% of Hispanic users of traditional medicine are not asked by their doctor whether they were using herbal remedies. Further, only one-third stated that herbs had potential interaction effects with prescription medicines. Thus, patients do not systematically reveal that they have been taking other medications, and this problem is particularly acute if large cultural differences (such as language barriers) exist between patient and doctor. Better levels of communication might be achieved if levels of trust between certain providers and individuals were higher (Russell 2005) and patients did not feel stigmatized for using traditional medicine. If physicians made it a standard practice to enquire about traditional medicine use, this may encourage reporting, increasing the chances of the individual’s full drug history being revealed. In turn, regulatory bodies need to start incorporating the potential side effects from drug interactions amongst more frequently used combinations, in order that physicians have such information to hand.

The study also sought to fill a gap in knowledge regarding outcome indicators for use of traditional medicines. Existing studies, at best, ask for the respondent’s level of satisfaction following treatment and do not use tangible measures such as those used in literature associated with measuring health outcome. The EQ5D has thus been applied to test the ‘before and after’ effect, and results show marked changes in quality of life and satisfaction, indicated by positive changes following traditional medicine use.

The high levels of satisfaction recorded suggest that users of traditional medicine/TMP are, on the whole, pleased with outcomes. Revealed high satisfaction scores show that in the short run traditional medicines are effective for relieving symptoms and improving health dimensions. This implies that even if access to orthodox medicines were improved, individuals may not necessarily switch to them, and consequently there may be some resistance to changes which fail to take into account the root reasons of traditional medicine utilization. These reasons may reflect cultural preference (the upshot being that orthodox and traditional medicines are not perfect substitutes) and genuine belief that traditional medicine is effective and sufficient. Policies which potentially increase access to orthodox medicines in developing countries—such as special pricing by pharmaceutical companies for developing countries—may find that individuals have already found their own solutions in dealing with common diseases like malaria, and uptake of modern medicines may not be as high as projected once traditional medicine use is factored in. Crucially, this suggests that policies must take into account the local context and environment, to research what systems prevail and be culturally sensitive in introducing new projects. Failure to do this neglects the fact that the existing institutional set-up has both sufficed and satisfied for hundreds of years.

Given its popularity, the question remains as to why people leave it to the second recourse, or a back up option, to try traditional medicine. Three explanations are offered here. Firstly, the revelation of traditional medicine use is closely guarded, especially where the presence of orthodox medical personnel is strong. People commented that doctors do not like, or think lowly of, traditional or herbal practices. Secondly, some do not see it as a ‘recourse’ at all because it is already part of their daily lives—many use traditional medicine as we might vitamins or as dietary supplements—so may fail to mention it. Thirdly, the use of medicines is known to be disease specific, so the use of traditional medicines is directly related to the symptoms experienced. All of these reasons could lead to under- or mis-reporting. Further investigations of disease or need standardised choices are required, in addition to an exploration of reasons for dissatisfaction with first recourse providers to partly answer these queries.

Findings show that an exploration into the key reasons for such high levels of satisfaction and the integration of the most popular elements could inform policy makers of potential changes to the orthodox system. Although the Ministry of Health has officially recognized the need to integrate traditional medicine, it remains silent on how this might be achieved from a users' perspective, leaning more towards the supply side, promoting regulation, quality and standards, working with healers and building scientific research and training capacity (MOHG 2007). Whilst all of these are vital, they would be better informed with indicators of utilization and resulting satisfaction. For example, individuals may be appreciative of the quality of care received from healers which leads to high levels of satisfaction. If this is so, policymakers need to reflect upon the care deficiencies in the orthodox system.

Traditional systems still freely coexist in conjunction with orthodox systems, and even where formal systems of care exist, knowledge about plant medicine and its preparations is strongly embedded within local communities and so use remains high. Demand for traditional medicine remains widespread and plays an important component of analysis once preferences and constraints are taken into account (Ensor and Cooper 2004). There is a definite role for both systems of care—and this study serves to strengthen the position of traditional medicine—but Ghana Health Service's goal of full integration is hampered by the lack of data which would give a clearer, more dynamic picture of the actions taken in sequelae by sick individuals, and the consequent levels of satisfaction. In this sense, longitudinal data tracking individuals over a prolonged period of time and a cost-benefit analysis using medicines cost data would add greatly to this study. Overall, the study advocates looking at the system holistically rather than traditional medicine being an appendage to the orthodox system. Policies and research which fail to do this will miss half the story when considering access to medicines in Ghana. In collating more demand-side data at the user level, policy making in resource-constrained countries would be better informed as to how to integrate traditional medicine with existing policies in a practical and fruitful manner.

Funding

This work was supported by partial funding from a scholarship provided by the London School of Economics, UK.

Conflict of interest

None declared.

Acknowledgements

I would like to thank Joan Costa-i-Font, Elias Mossialos, Ben Baumberg, Francesco D'Amico, colleagues in Ghana and two anonymous reviewers for their valuable comments and support.

References

Aday
LA
Andersen
RA
A framework for the study of access to medical care
Health Services Research
 , 
1974
, vol. 
9
 (pg. 
208
-
20
)
Anyinam
C
Availability, accessibility, acceptability and adaptability: four attributes of African ethno-medicine
Social Science & Medicine
 , 
1987
, vol. 
25
 (pg. 
303
-
11
)
Brieger
W
Osamor
P
Salami
K
Oladepo
O
Otusanya
S
Interactions between patent medicine vendors and customers in urban and rural Nigeria
Health Policy and Planning
 , 
2004
, vol. 
19
 (pg. 
177
-
82
)
de-Graft Aikins
A
Healer-shopping in Africa: new evidence from a rural-urban qualitative study of Ghanaian diabetes experiences
British Medical Journal
 , 
2005
, vol. 
331
 pg. 
737
 
Dolan
P
Lee
H
Peasgood
T
Losing sight of the wood for the trees: some issues in describing and valuing health and another possible approach
Pharmacoeconomics
 , 
forthcoming
Ensor
T
Cooper
S
Overcoming barriers to health service access: influencing the demand side
Health Policy and Planning
 , 
2004
, vol. 
19
 (pg. 
69
-
79
)
Flatie
T
Gedif
T
Asres
K
Gebre-Mariam
T
Ethnomedical survey of Berta ethnic group Assosa Zone, Benishangul-Gumuz regional state, mid-west Ethiopia
Journal of Ethnobiology and Ethnomedicine
 , 
2009
, vol. 
5
 pg. 
14
 
Goodman
C
Brieger
W
Unwin
A
, et al.  . 
Medicine sellers and malaria treatment in Sub-Saharan Africa: what do they do and how can their practice be improved?
American Journal of Tropical Medicine and Hygiene
 , 
2007
, vol. 
77
 
6 Suppl.
(pg. 
203
-
18
)
Howell
L
Kochhar
K
Saywell
R
, et al.  . 
Use of herbal remedies by Hispanic patients: do they inform their physician?
Journal of the American Board of Family Medicine
 , 
2006
, vol. 
19
 (pg. 
566
-
78
)
Kind
P
Dolan
P
Gudex
C
William
A
Variations in population health status: results from a United Kingdom national questionnaire survey
British Medical Journal
 , 
1998
, vol. 
316
 (pg. 
736
-
41
)
Kroeger
A
Anthropological and socio-medical health care research in developing countries
Social Science & Medicine
 , 
1983
, vol. 
17
 (pg. 
147
-
61
)
Meyer-Weitz
A
Reddy
P
Van den Borne
HW
Kokd
G
Pietersene
J
Health care seeking behaviour of patients with sexually transmitted diseases: determinants of delay behaviour
Patient Education and Counselling
 , 
2000
, vol. 
41
 (pg. 
263
-
74
)
Mills
E
Cooper
C
Seely
D
Kamfer
I
African herbal medicines in the treatment of HIV: Hypoxis and Sutherlandia. An overview of evidence and pharmacology
Nutrition Journal
 , 
2005
, vol. 
4
 pg. 
19
 
Ministry of Health, Ghana (MOHG)
Ghana National Drugs Policy
2004
Second edition. Accra, Ghana
 
Ministry of Health, Ghana (MOHG)
Creating wealth through health: the health sector programme of work
2007
 
2007–2011. Online at: http://www.moh-ghana.org/UploadFiles/Policies_Programs/5YPoW (2007-2011)100106035321.pdf (Accessed June 2011)
Mshana
G
Hampshire
K
Panter-Brick
C
Walker
R
Urban-rural contrasts in explanatory models and treatment-seeking behaviours for stroke in Tanzania
Journal of Biosocial Science
 , 
2008
, vol. 
40
 (pg. 
35
-
52
)
Mwabu
G
Health care decisions at the household level: results of a rural health survey in Kenya
Social Science & Medicine
 , 
1986
, vol. 
22
 (pg. 
315
-
9
)
Peltzer
K
Community perceptions of biomedical health care in a rural area in the Northern Province, South Africa
Health SA Gesondheid
 , 
2000
, vol. 
5
 
1
Peltzer
K
Friend-du Preez
N
Ramlagan
S
Fomundam
H
Use of traditional complementary and alternative medicine for HIV patients in KwaZulu-Natal, South Africa
BMC Public Health
 , 
2010
, vol. 
8
 pg. 
255
 
Russell
S
Treatment seeking behaviour in urban Sri Lanka: trusting the state, trusting private providers
Social Science & Medicine
 , 
2005
, vol. 
61
 (pg. 
1396
-
407
)
Sachs
L
Tomson
G
Medicines and culture—a double perspective on drug utilization in a developing country
Social Science & Medicine
 , 
1992
, vol. 
34
 (pg. 
307
-
15
)
Sizmur
S
Redding
D
Core domains for measuring inpatients’ experience of care. Picker Institute discussion paper
2009
 
Smith
F
Community pharmacy in Ghana: enhancing the contribution to primary health care
Health Policy and Planning
 , 
2004
, vol. 
19
 (pg. 
234
-
41
)
Stekelenburg
J
Jager
BE
Kolk
PR
, et al.  . 
Health care seeking behaviour and utilisation of traditional healers in Kalabo, Zambia
Health Policy
 , 
2005
, vol. 
71
 (pg. 
67
-
81
)
Szende
A
Oppe
M
Devlin
N
EQ5D Value Sets: Inventory, Comparative Review and User Guide
 , 
2007
 
EuroQol Group Monographs, Vol 2. Dordrecht: Springer
Tabi
MM
Powell
M
Hodnicki
D
Use of traditional healers and modern medicine in Ghana
International Nursing Review
 , 
2006
, vol. 
53
 (pg. 
52
-
8
)
Tsey
K
Traditional medicine in contemporary Ghana: a public policy analysis
Social Science & Medicine
 , 
1997
, vol. 
45
 (pg. 
1065
-
74
)
Twumasi
PA
A social history of the Ghanaian pluralistic medical system
Social Science & Medicine, Medical Anthropology
 , 
1979
, vol. 
13
 (pg. 
349
-
56
)
UN Millennium Project
Prescription for Healthy Development: Increasing Access to Medicines. Report of the Task Force on HIV/AIDS, Malaria, TB, and Access to Essential Medicines, Working Group on Access to Essential Medicines
 , 
2005
Sterling, VA
Earthscan
van den Boom
GJM
Nsowah-Nuamah
NNN
Overbosch
GB
Health care provision and self-medication in Ghana. Paper
2004
 
van der Geest
S
Whyte
S
The Context of Medicines in Developing Countries: Studies in Pharmaceutical Anthropology
 , 
1988
Dordrecht
Springer
WHO
WHO Traditional Medicine Strategy 2002–2005
 , 
2002
 
WHO
Medicines definitions
2011
 
Winslow
LC
Kroll
DJ
Herbs as medicines
Archives of Internal Medicine
 , 
1998
, vol. 
158
 (pg. 
2192
-
9
)

Appendix

Table A1

Total number of recourses to care sought by acute and chronic illness and region, weighted

 No. with needs Number of recourses sought
 
1+ incl. SM 1+ excl. SM 2+ incl. SM 2+ excl. SM 3+ incl. SM 3+ excl. SM 4+ incl. SM 4+ excl. SM 
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) 
Acute 
    GA 381 4 (1.0) 377 (99.0) 331 (86.9) 87 (22.8) 68 (17.8) 11 (2.9) 7 (1.8) 0 (0.0) 0 (0.0) 
    UW 60 3 (5.0) 57 (95.0) 53 (88.3) 18 (30.0) 13 (21.7) 2 (3.3) 1 (1.7) 0 (0.0) 0 (0.0) 
    Total 441 7 (1.6) 434 (98.4) 384 (87.1) 105 (23.8) 81 (18.4) 13 (2.9) 8 (1.8) 0 (0.0) 0 (0.0) 
Chronic 
    GA 359 2 (0.6) 357 (99.4) 347 (96.7) 126 (35.1) 96 (26.7) 16 (4.5) 16 (4.5) 7 (1.9) 0 (0.0) 
    UW 27 1 (3.7) 26 (96.3) 26 (96.3) 17 (63.0) 17 (63.0) 7 (25.9) 7 (25.9) 1 (3.7) 0 (0.0) 
    Total 386 3 (0.8) 383 (99.2) 373 (96.6) 143 (37.0) 113 (29.3) 23 (6.0) 23 (6.0) 8 (2.1) 0 (0.0) 
 No. with needs Number of recourses sought
 
1+ incl. SM 1+ excl. SM 2+ incl. SM 2+ excl. SM 3+ incl. SM 3+ excl. SM 4+ incl. SM 4+ excl. SM 
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) 
Acute 
    GA 381 4 (1.0) 377 (99.0) 331 (86.9) 87 (22.8) 68 (17.8) 11 (2.9) 7 (1.8) 0 (0.0) 0 (0.0) 
    UW 60 3 (5.0) 57 (95.0) 53 (88.3) 18 (30.0) 13 (21.7) 2 (3.3) 1 (1.7) 0 (0.0) 0 (0.0) 
    Total 441 7 (1.6) 434 (98.4) 384 (87.1) 105 (23.8) 81 (18.4) 13 (2.9) 8 (1.8) 0 (0.0) 0 (0.0) 
Chronic 
    GA 359 2 (0.6) 357 (99.4) 347 (96.7) 126 (35.1) 96 (26.7) 16 (4.5) 16 (4.5) 7 (1.9) 0 (0.0) 
    UW 27 1 (3.7) 26 (96.3) 26 (96.3) 17 (63.0) 17 (63.0) 7 (25.9) 7 (25.9) 1 (3.7) 0 (0.0) 
    Total 386 3 (0.8) 383 (99.2) 373 (96.6) 143 (37.0) 113 (29.3) 23 (6.0) 23 (6.0) 8 (2.1) 0 (0.0) 

Source: author's own

Notes: Derivation of weights: weight using gender statistics by region, for 2010:

1. Apportion percentages of total for each of the following using population data: male-Upper West, male-Greater Accra, female-Upper West, female-Greater Accra.

2. Compare each category with sample data.

3. Derive weights by calculating population statistics/sample statistics.

4. Weights are: male-Upper West 0.222445, male-Greater Accra 2.363463, female-Upper West 0.216964, female-Greater Accra 1.86199.

SM indicates self-medication, percentages given as proportion of row total. GA = Greater Accra. UW = Upper West.

Figures rounded to nearest whole number.

Table A2

Changes in dimension, by type of illness, unweighted

 Weighted changes in dimension
 
 Acute +  Chronic Acute Chronic 
Mobility 
    −2 0.2 0.0 0.0 
    −1 0.1 0.0 0.0 
    0 97 44.8 44 46.3 60 44.4 
    1 92 42.3 37 38.9 61 45.2 
    2 27 12.5 14 14.7 14 10.4 
    Total 217 100 95 100 135 100 
Self care 
    −2 0.2 0.0 0.0 
    −1 0.1 0.0 0.0 
    0 99 45.7 47 50.0 58 43.0 
    1 89 40.9 34 36.2 57 42.2 
    2 28 13.1 13 13.8 20 14.8 
    Total 217 100 94 100 135 100 
Activities 
    −2 1.2 2.1 0.0 
    −1 0.2 0.0 0.0 
    0 87 40.2 37 39.4 54 40.3 
    1 96 44.0 40 42.6 64 47.8 
    2 31 14.4 15 16.0 16 11.9 
    Total 217 100 94 100 134 100 
Pain 
    −2 1.2 2.1 0.0 
    −1 0.2 0.0 0.0 
    0 48 22.1 16 17.0 32 23.7 
    1 100 46.2 43 45.7 64 47.4 
    2 66 30.3 33 35.1 39 28.9 
    Total 217 100 94 100 135 100 
Anxiety 
    −2 0.2 0.0 0.0 
    −1 2.1 2.1 1.5 
    0 93 43.2 48 51.1 49 36.8 
    1 95 43.9 35 37.2 67 50.4 
    2 23 10.6 9.6 15 11.3 
    Total 215 100 94 100 133 100 
 Weighted changes in dimension
 
 Acute +  Chronic Acute Chronic 
Mobility 
    −2 0.2 0.0 0.0 
    −1 0.1 0.0 0.0 
    0 97 44.8 44 46.3 60 44.4 
    1 92 42.3 37 38.9 61 45.2 
    2 27 12.5 14 14.7 14 10.4 
    Total 217 100 95 100 135 100 
Self care 
    −2 0.2 0.0 0.0 
    −1 0.1 0.0 0.0 
    0 99 45.7 47 50.0 58 43.0 
    1 89 40.9 34 36.2 57 42.2 
    2 28 13.1 13 13.8 20 14.8 
    Total 217 100 94 100 135 100 
Activities 
    −2 1.2 2.1 0.0 
    −1 0.2 0.0 0.0 
    0 87 40.2 37 39.4 54 40.3 
    1 96 44.0 40 42.6 64 47.8 
    2 31 14.4 15 16.0 16 11.9 
    Total 217 100 94 100 134 100 
Pain 
    −2 1.2 2.1 0.0 
    −1 0.2 0.0 0.0 
    0 48 22.1 16 17.0 32 23.7 
    1 100 46.2 43 45.7 64 47.4 
    2 66 30.3 33 35.1 39 28.9 
    Total 217 100 94 100 135 100 
Anxiety 
    −2 0.2 0.0 0.0 
    −1 2.1 2.1 1.5 
    0 93 43.2 48 51.1 49 36.8 
    1 95 43.9 35 37.2 67 50.4 
    2 23 10.6 9.6 15 11.3 
    Total 215 100 94 100 133 100 

Endnotes

1 It must be noted that whilst the aim was to achieve an even spread of facilities and households within a given district, there is no guarantee of no overlap of facilities within a radius. Thus a household may have several facilities close by, particularly in more urban areas more densely populated by facilities and people. However, reference facilities were seen as a starting point for sampling rather than a pure indicator of access, especially given that numerous other sources of care (including self-medication at home, use of informal providers in the community and so on) were also available to individuals, and therefore this was not seen as an impediment to sampling.
2 Additionally, where possible, the respondent in question was called upon to answer their individual questionnaire modules.
3 At the time of writing, the most recent census figures (2010) were not yet available.
4Appendix Table 1 shows weighted data which do not change the main result of the study, as the percentages of individuals using certain numbers of recourses do not differ significantly. For example, the majority of individuals still use at least one source of care, there is a greater likelihood of using multiple sources when suffering from chronic illnesses, and almost nobody uses more than three sources. When the data are broken down by region, the proportion of people expected to utilize different numbers of care sources remains fairly steady, although raw counts change according to the assigned weight.
5 A weighted version of the table is given in Appendix Table 2 using the same weights as Appendix Table 1. Again, the general trend remains unchanged, with a large majority of people reporting high satisfaction scores, with many positive and very few negative changes indicated.