Introduction

Diabetes and hypertension are the highest burden of non-communicable diseases to populations worldwide [1, 2]. An estimated 366 million and 1 billion people globally are living with diabetes and hypertension respectively [3, 4]. By 2020, prevalence of these diseases is expected to increase between 13% and 30% [1, 2], without more innovative preventive interventions. Diabetes and hypertension prevalence are increasingly high in Low-and Middle- Income-Countries (LMICs) that are experiencing rapid transition in health and diseases patterns, and their interaction with the social determinants of health [5]. Governments in LMICs countries have over the years increased investment on interventions to improve health systems responsiveness as a vehicle to control and reduce the health, social and economic burden of these non-communicable diseases [6, 7]. Such investments, however, are undermined by persistent unhealthy lifestyle choices including tobacco use, poor diet and physical inactivity [5, 8].

Awareness of risk factors and preventive measures for diabetes and hypertension is low in LMICs compared to high income regions [8, 9]. In some countries, physical and financial access to appropriate diagnostics for high risk populations is a challenge as is treatment medication for affected individuals [10, 11]. Even if individuals are diagnosed and admitted to long-term drug therapy essential for managing diabetes and hypertension, poor adherence to prescription often pose as a barrier to realising effective therapeutic outcomes [12].

Adherence to drug therapy describe the extent of a person’s behaviour in terms of taking drugs or executing life style changes in conformity with health provider recommendations [13]. Non-adherence arises when a patient irregularly follow prescription or discontinue use [14], for reasons either intended or unintended [15]. Non-adherence often increases the risk of relapse, poor therapeutic outcomes and needless mortalities [14]. Hospital admissions and readmissions, emergency outpatient visits and mortalities increases when patients poorly adhere to medication as directed [16, 17]. Increase hospital visits due to non-adherence in turn contributes to depletion of available scarce resources and supplies in health facilities [18].

Prevalence of medication non-adherence is on the ascendancy especially in LMICs where adherence averages less than 50% [13]. In Ghana as with many other LMICs, national level data on medication adherence particularly among diabetes and hypertension patients is sparse. Studies in two different settings, however, found 38.5% adherence rate among diabetes patients [19] and 47.7% for hypertensive patients [20]. These statistics are unacceptably lower than the recommended 90% threshold [21]. This makes it urgent and compelling to uncover the underlying factors influencing poor medication taking behaviour among these patient cohorts. Yet, studies drawing policy makers and practitioners’ attention to the wide ranging factors influencing non-adherence remains limited in Ghana.

A handful of quantitative studies have identified age, income, education and gender as demographic factors influencing non-adherence to malaria and diabetic drug treatment [19]. Such demographic analysis are crucial, but they tend to limit understanding of the complex set of factors that on their own or in interaction with others influence non-adherence in terms of skipping doses or discontinuation of meditation taking [22]. Also missing in existing studies is why non-adherence occur among hypertensive patients. To fill this research gap, this study qualitatively explored how and why chronic diabetic and hypertensive patients admitted to long-term drug therapy poorly adhered to prescribed medication taking.

Methods

The study was conducted in the Korle-Bu Teaching Hospital located in Ghana’s capital, Accra, in the Greater Accra region. Starting from less than 200 beds and 200 outpatient visits at its inception in 1923 [23], the hospital has seen tremendous expansion over the years with about 2000 beds, 21 clinical and diagnostic departments and 3 Centres of Excellence. Average daily outpatient visits and inpatient admissions have increased to about 1500 and 250 respectively [24]. Korle Bu is the top referral point for patients within and across Ghana, especially West African Countries [25]. Chronic communicable and non-communicable diseases are often among the highest reported cases and associated mortalities in the hospital [26].

We targeted patients admitted to long-term diabetes and hypertension drug therapy in the hospital, but were subsequently non-adherent, either by skipping doses or discontinued with the medication taking [12], leading to relapse and readmission. In order to cross validate patients’ perspectives and also account for their limitations in providing narratives in more detail, we included their caregivers at the time of admission in the hospital. The study was qualitative, employing in-depth interviews using a discussion guide to explore factors that pertained but not limited to personal, family or community and health provider factors that may have influenced non-adherence. These factors were derived from a cursory review of existing literature on medication non-adherence [27, 28].

Interviews were held separately for patients and their caregivers during admission and/or the day of discharge depending on their request. Interviews took the form of discussions, using series of inductive probes to widen the narratives as much as possible. This process enabled the gathering of additional insightful factors not covered in the discussion guide. Notes were taken during interviews and if participants agreed they were audiotaped. Interviews were conducted in either English or Twi (a local language predominantly spoken in every part of Ghana) depending on participants choosing and lasted for an average of about 1 hour.

The team of data collectors comprised two of the authors (RAA and NMG) and one research assistant who was given a day’s training on qualitative methods. The team visited the hospital’s medical ward biweekly, each time being assisted by the ward nurses to purposively recruit participants into the study. In the first month of data collection, 4 patients and 2 caregivers were recruited. More or equal number of participants were subsequently recruited till thematic variations were observed [29]. At the end of the study period–July to December 2015, participants totalled 49; comprising 32 patients and 17 caregivers (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Numbers of participants recruited each month, July and December 2015.

https://doi.org/10.1371/journal.pone.0193995.t001

The study received approval by the Ghana Health Service. Verbal consent was sought from each participant. Prior to each interview, the study’s purpose and its potential contribution to improving medication adherence were explained. Next, patients and their caregivers were made aware that participation in the study was voluntary, with a further assurance of their right to withdraw from the interview at any time.

Audio recordings in the local language (Twi) were translated into English and crossed verified by the last author (NMG) who simultaneously compared the audios and transcripts. The transcripts together with the field notes were read thoroughly, refined and then exported into Nvivo version 11 for analysis by RAA (the first Author). To ensure rigour and accuracy of the software coding process, the transcripts were doubled coded by RAA and an independent person guided by Attride-Stirling [30] stages of thematic network analysis. In the first stage, we set out to breakdown the text by assigning codes to key words or phrases reflecting participants’ viewpoints on non-adherence–discontinuation of medication-taking, missing of doses or both. About 102 codes were derived at this stage based on their frequency of occurrence, but only 35 are reported. The second stage extracted basic themes briefly describing the codes generated from stage one. Basic themes were then represented by organising themes in the third stage. Finally, global themes were generated and agreed upon to give an overall meaning to the lower-order themes abstracted [30]. Table 2 illustrates the codes and themes building process. Results are presented based on the broad themes and verbatim quotes referenced as either hypertensive or diabetic patient, or comorbidity patient (those with both conditions).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Codes and themes building process.

https://doi.org/10.1371/journal.pone.0193995.t002

Results

Participants’ characteristics

Table 3 shows that the majority of the participants were males (57.1%), aged 40 years and above (60.7%), employed (39.3%), Christians (53.6%) and currently married (64.3%). There was slightly more diabetes (39.3%) than hypertension (32.1%) patients, while 28.6% were suffering from both conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Characteristics of study participants.

https://doi.org/10.1371/journal.pone.0193995.t003

Discussion

We found non-adherence was the result of patients ill perception of medications efficacy, a finding that parallels others [28, 31], but contrasts with another involving patients from, and across Ghana, that showed perceived medication efficacy influenced non-adherence [32]. Patients also doubted efficacy of the medications and therefore did not have patience to follow the dosage till completion. We attribute this to common medication-taking behaviour in Ghana. A related study by Dodor and Afenyadub [33], showed that treatment default among TB patients had to with ill-feeling about drugs and the impatience to routinely take prescribed medication till completion. Such behaviours are usually the cause of poor outcomes leading to patients engaging in ‘healer shopping’ [34]. This finding suggest the need for patients to be properly counselled during dispensing to enable them appreciate the proven scientific efficacy of medications in the management of diabetes and hypertension. Health providers must stress that the benefits of medications are realised when patients carefully and regularly follow dosage regime.

The behaviour of patients refusing medication and turning to traditional herbal medicine is a growing phenomenon impeding orthodox care outcomes in Ghana [35]. Limited trust in, and perceived inefficacy of modern care usually account for such behaviours. A study in rural and urban Ghana, for example, showed that acute and chronic patients first sought modern care, but turned to traditional herbal medicine because of perception that modern medication is not effective [35]. The danger about herbal medicine is that their efficacy and toxicity has not been tested scientifically. Thus, users of traditional herbal medications are prone to developing viral resistance to medications and experiencing complications [36], as manifested in the results. The finding about patients resorting to traditional healers to revoke unnatural causes relating to sorcery and spells complements other studies in Ghana [37, 38], and suggests the need for interventions to pursue patients to rely solely on biomedical approaches for managing conditions. One such intervention is the suggestion by Aikins [34] that illnesses caused by belief systems should be treated with ethnomedical drugs scientifically approved to be nontoxigenic. Medical and ethnomedical practitioners can then collaborate to treat symptoms when necessary.

Non-adherence was associated with spiritual or divine healing through prayers, fasting and paranormal therapies. This findings reflects in the growing demand for spiritual healers services in treating conditions among segments of populations in Ghana [39, 40]. Patients usually prefer spiritual or divine healing because of its affordability and accessibility [41] as well as the belief that certain illnesses have spiritual causations beyond medical treatment [38]. The problem about spiritual or divine healing is the conflict with modern care [42]. This was brought to light as spiritual healers discouraged patients from taking the medications because it was incompatible with divine healing ways [41]. The belief that medication is unnecessary because ill-health is an act of God and uncontrollable except divine intervention supports Tabi et al. [38], whose study revealed individuals preferred healing by God to be the most safest.

The practice of polypharmacy leading to complications is a disturbing finding. In one breath, it reflects in irrational consumption of medications as a quick fix, when in fact the prescribed medications sufficed. On the other hand, it reflects in poor education given by providers on the consequences of consuming non-prescribed drugs while taking the medication. Whatever the problem, the fact remains that polypharmacy practice caused hazardous drug-drug and drug-disease interactions [43], resulting in patients discontinuing the medication-taking. This is in keeping with the finding that non-adherence increases with increasing consumption of more non-prescribed medications [44]. To avoid the risk of polypharmacy practice, patients and their caregivers should be educated on drug safety, and the importance of sticking to prescribed drug regime. Adverse drug effect due to polypharmacy should be explained and caution given to patients to report symptoms, rather than resort to self-medication with over-the-counter-drugs.

Cognitive problems relating to forgetfulness emerged as a barrier of adherence. Forgetfulness had underlying causes: routine work schedules and frequent travels, both of which disrupted daily medication schedule [45]. Among the factors that accounted for non-adherence, forgetfulness appears to be the most difficult to control. Because, its occurrence is involuntary. Despite this, we propose that patients exhibiting forgetfulness be given cognitive support. This can take the form of family members using cues to remind patients of daily medication schedules. Patients can also benefit from adherence through text messaging or phone call reminders from social and work peers. The use of alarms and pill boxes by patients prone to forgetfulness is also advised [17].

Findings showed that societal beliefs and norms shaped patients’ decision to refrain from adherence, because they are at variance with biomedical therapies for illnesses. The cultural path dependency theory provide insight into the phenomenon of norms shaping medication-taking [46]. The theory states that individuals although independent are governed by group norms and feedbacks from the wider society. Thus, societal socio-cultural norms that ‘incorrectly’ prohibit modern care is a feedback that can change individuals own actions to approximate the social environment [47]. We suggest the need for health providers to adopt cultural competence therapeutic approaches that take into account patients’ beliefs, values and norms in administering medications [48]. Knowing the patient’s ethnographic profile including prevailing norms can be useful in providing constructive prescription counselling.

Health providers’ actions were shown to influence non-adherence. They failed to properly explain the dosage regime to the understanding of patients. Also, a patient-centred approach of involving patients in prescribing decisions and answering pertinent questions of concern was lacking. This worked out to compromise adherence. Indeed, medication use is not only limited to telling the patient how and when to take drugs, but also, supporting the patient to take drugs by considering her preferences, beliefs and concerns in making prescribing decisions [22]. Another provider side factor that influenced non-adherence was poor written instructions about the dosage regime. Patients admitted to being confused as they tried to interpret the dosage regime due to poor hand writing. To get around this, they devised their own dosage calculation which was incorrect leading to complications and discontinuation of the medication [49]. We suggest the need for health providers, particularly prescribers to improve on their writing quality such that it convey simple information across to the patient.

In summary, findings showed how and why non-adherence to diabetes and hypertension medication is variously influenced by all the eight factors. However, a closer look at the results shows that three of the factors, namely perceived medication inefficacy; recourse to herbal medicine and recourse to spiritual or divine healing strongly interacted to influence non-adherence. Thus, we drew on a causal loop diagram (CLD) (Fig 1), using Vensim software [50], to visualise and explain the interaction effect and feedback mechanisms of the three factors influence on non-adherence. As the Fig 1 shows, the more patients doubted the medication efficacy, the more they preferred herbal medicine, which in turn increases the perception that diabetes and hypertension medications do not have effective therapeutic effect. This reinforcing effect is illustrated by Loop 1. Similar reinforcing effect can be said of Loop 2 where ill perception of medication efficacy adds to the growing problem of recourse to spiritual or divine healing with a reversible consequence of increasing the burden of poor perception of the medications efficacy. In Loop 3, the more patients seek spiritual healing, the more likely is it that non-adherence increases which in turn produces a reinforcing effect on resource to spiritual healing. The same explanation can be given to Loop 4. Overall, recourses to herbal and spiritual healing ways increase with increasing ill-perception of the medications efficacy, all of which combine to strengthen non-adherence.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CLD of the interaction effect of perceived medication efficacy, recourses to herbal medicine and spiritual healing on non-adherence to diabetes and hypertension medication.

https://doi.org/10.1371/journal.pone.0193995.g001

Acknowledgments

We thank the patients and caregivers for agreeing to participate in the study. We are also indebted to the health providers for their immense assistance.