Cognitive predictors of adherence to malaria prophylaxis regimens on return from a malarious region: a prospective study

Abstract

Cases of `imported malaria' into countries where malaria is not endemic are increasing and evidence suggests that non-use of malaria prophylaxis and lack of adherence are contributing to this increase. Non-adherence may be especially likely because chemoprophylaxis regimens require travellers to continue to take medication for 4 weeks after their return from a malarious region. This study investigated the extent to which cognition measures specified by the theory of planned behaviour and the health belief model could distinguish between those who reported greater or lesser adherence after their return. Cognitions were measured using a brief questionnaire on the day of departure from the malarious region and reports of adherence were collected between 5 and 7 weeks later. Data from two longitudinal samples of UK tourists returning from The Gambia were analysed; 106 mefloquine users and 61 chloroquine and proguanil users. Results suggested that malaria prophylaxis adherence could be improved. 22.5% of mefloquine users and 31% of chloroquine and proguanil users reported adherence for 3 weeks or less. A model based on the theory of planned behaviour explained approximately 50% of the variance in reported adherence amongst mefloquine users and 40% amongst chloroquine and proguanil users, comparing favorably with other published applications of the theory. Findings suggest that targeting key cognitions could enhance adherence on return from malarious regions. Enhancing perceived control over adherence may be important as well as emphasising susceptibility to malaria infection. Reassuring mefloquine users concerning potential side effects of the drug may also encourage adherence on return. Implications for future research are discussed.

Introduction

With the continued expansion of international tourism, new markets have developed in travel to destinations in `under-developed' countries in tropical regions. Visitors to such countries are exposed to a range of health risks and diseases, they do not encounter at home. Foremost amongst these, are the various forms of malaria (Behrens and Curtis, 1993). Unlike many tropical diseases, immunisation by vaccine is not yet available but various chemoprophylaxis regimens reduce the risk of contracting the disease (Behrens, 1997). Careful adherence to prophylactic regimes is essential to ensure maximal protection and medication must be taken for 4 weeks after returning from a malarious region because the infection may persist in the body for that period.

Global increases in malaria, the emergence of drug resistance and the greater volume of travellers into malarious regions, have combined to produce increasing numbers of cases of `imported malaria' into countries where malaria is not endemic. Surveillance of imported malaria is undertaken across western Europe but available figures are likely to under-estimate the problem (Legros et al., 1998). Legros et al., for example, suggest that registered cases in France may represent less than half of the actual number. Nevertheless, available data indicates clearly that the scale of the problem is increasing, that most cases of the potentially fatal falciparum malaria were acquired during visits to countries in sub-Saharan Africa and that large proportions of people succumbing to malaria either failed to take chemoprophylactic measures or did not adhere adequately to their prescribed regimen. Non-use of malaria prophylaxis or lack of adherence is, therefore, a crucial factor associated with imported malaria in western European countries and increased efforts to improve use of malaria prophylaxis are needed and can be readily justified on economic grounds (Behrens and Roberts, 1994).

In the UK, for example, 2500 cases were recorded in 1996, the highest annual total for 25 years; 51% of cases were the falciparum type; almost all cases of malaria, with the exception of vivax malaria were associated with visits to Africa and for cases of falciparum malaria, 64% of affected individuals took no form of prophylaxis (Bradley et al., 1998). In Italy, the UK and the Netherlands imported malaria is recognised to be a particular and increasing problem among immigrant groups (Wetsteyn et al., 1997) while, outside Europe, imported malaria is a problem for all affluent non-malarious countries with internationally mobile populations (e.g. New Zealand (Kriechbaum and Baker, 1996) and Japan (Kimura et al., 1996).

Studies of malaria prophylaxis amongst general samples of tourists traveling to malarious regions could provide guidance on the promotion of greater preventive efforts. Most studies to date have employed retrospective surveys of returned travellers Held et al., 1994, Gagneux et al., 1996. In one of the few longitudinal studies available, Huzly et al. (1996) investigated adherence amongst more than 6,500 German tourists recruited during pre-travel consultations. Telephone interviews were conducted 4 weeks after return from travel and indicated an overall adherence rate of 83% with better adherence for mefloquine than other regimens. Prophylaxis was discontinued or taken irregularly mainly during the 4 weeks after return. Key reasons given for non-adherence were side effects, forgetfulness, a lack of perceived risk and discomfort when swallowing tablets. Additional factors found to be significantly associated with non-adherence were: business travel, not taking tablets at a regular time or with meals, staying mainly in cities, experience of side effects previously, taking other drugs while away and stays of less than a week or more than 4 weeks.

A number of social cognitive models have been developed to identify potentially modifiable beliefs, attitudes and intentions which predict health behaviours (see Conner and Norman, 1996, Abraham and Sheeran, 1997 for reviews) The most widely applied, the `health belief model' (HBM) (Becker et al., 1977), highlights two aspects of individual representations of health behaviour; threat perception and behavioral evaluation. Each consists of two distinct sets of beliefs. Threat perception involves perceived susceptibility to illness and anticipated severity of the consequences of the illness. Behavioral evaluation depends upon beliefs concerning the benefits or efficacy of a recommended health behaviour and those concerning the costs of enacting the behaviour.

The utility of the HBM in characterising those who do and do not adhere to medicine regimens has been tested across different patient groups Bloom-Cerkoney and Hart, 1980, Cummings et al., 1981, Hogan et al., 1983, Brownlee-Duffeck et al., 1987, Kelly et al., 1987. However, Harrison et al. (1992) were only able to include two prospective studies of adherence in their meta-analytic review. Reid and Christensen (1988) found that the HBM was able to explain 10% of the variance in adherence amongst women taking tablets for urinary tract infections, Kelly et al. (1987) found that the model explained 20% of the variance in reported compliance amongst psychiatric out-patients, while Brownlee-Duffeck et al. (1987) found that it accounted for 40% of the variance in reported compliance amongst insulin-dependent diabetics. These are encouraging results because they suggest that modifiable beliefs may influence adherence behaviour. For example, using their `binomial effect size display', Rosenthal and Rubin (1982) calculate that explaining 20% of variance corresponds to a potential success-rate increase of 44%. In other words, if 28% of participants in a control group perform a health behaviour, we might expect 72% of those who had adopted the relevant beliefs to undertake the target behaviour.

Further research on the cognitive antecedents of health-related behaviour has indicated that other models, such as the theory of reasoned action (TRA) can account for greater proportions of the variance in behaviour than the HBM Conner and Norman, 1994, Zimmerman and Vernberg, 1994, Sheeran and Abraham, 1996. For example, Reid and Christensen (1988) observed that the HBM explained 10% of the variance in adherence to a tablet regimen, but the variance accounted for increased to 29% when cognitions specified by the theory of reasoned action were added.

The TRA (Fishbein and Ajzen, 1975) proposes a cognitive mechanism by which beliefs about the threat of illness and preventive behaviour are translated into action, i.e. intention formation. Intention formation is thought to precede and predict behaviour and to be determined by beliefs constituting a person's attitude and subjective norm regarding a particular behaviour. Attitudes are based on beliefs concerning the likely consequences of a particular behaviour and evaluations of those consequences. These include the beliefs specified by the HBM. Subjective norm measures combine perceptions of the extent to which other people approve of a behaviour and the strength of motivation to comply with these people's wishes.

Bandura, 1992, Bandura, 1997 has argued that perceived self-efficacy, that is the belief that one can successfully perform a specified behaviour, affects motivation and behavioural control. In general, those who believe they will succeed are more likely to formulate intentions to act, set higher performance standards, exert greater effort, regard errors as learning experiences and persevere for longer. The importance of self-efficacy beliefs has been acknowledged by their inclusion in a reformulation of the HBM (Rosenstock et al., 1988) and the addition of a very similar concept to the theory of reasoned action. Ajzen and Madden (1986) tested a revised TRA (the theory of planned behaviour, TPB), which included a measure of perceived behavioural control which was found to have direct effects on behaviour as well as indirect effects through its relationship with intention formation (Ajzen, 1991).

The TRA and the TPB have been used to identify cognitive antecedents of a range of health-related behaviours Conner and Sparks, 1996, Godin and Kok, 1996 but there have been relatively few tests of these models and particularly the potentially more powerful TPB, in relation to medical adherence (e.g. Cochran and Gitlin, 1988, Reid and Christensen, 1988, Miller et al., 1992). In one of the few tests of the TPB, Hounsa et al. (1993) found that the theory was able to explain 40% of the variance in mothers' intention to use oral rehydration therapy but adherence itself was not measured. No applications of these models to malaria prophylaxis were found during the present review.

The aim of this study was to assess the extent to which a cognitive model of adherence based on the TPB and the HBM was able to specify cognitions which could distinguish between those who do and do not adhere to malaria prophylaxis regimens. Adherence after return from a malarious region was investigated as this has been identified as a period when travellers may be less likely to take medication as directed (Huzly et al., 1996), perhaps due to a reduced sense of risk.

Previous findings have indicated that elements of the HBM are important correlates of adherence to medical regimens and it has been suggested that these beliefs should be measured separately rather than within general attitude measures (Schwarzer, 1992). Consequently, the model used here combined assessments of perceived susceptibility, perceived severity and perceived side effects (which has been found to be the main cost of malaria prophylaxis) with measures specified by the TPB. This is in line with the Ajzen (1998) recent suggestion that extra components can be added to the model where these are found to explain addition variance in particular behaviours.

The present study extended previous survey work on UK travellers to The Gambia. In the UK, a specialist committee within the Public Health Laboratory Service is responsible for issuing guidance on malaria prophylaxis. A range of anti-malaria drugs is available (Behrens, 1997) and the choice of drug should be based on the risk of malaria, the level of drug resistance in the country to be visited and the person's medical history. At the time of this study, the risk of transmission of malaria in The Gambia was high and malaria was chloroquine resistant. Mefloquine was the drug of first recommendation, unless contra-indicated on medical grounds. For those not taking mefloquine, a dual regimen of chloroquine plus proguanil was recommended. Travellers using mefloquine were advised to take one tablet a week while those using the dual regimen, were advised to take a dose of proguanil every day and a dose of chloroquine weekly. In both cases it was recommended that medication be taken for at least 1 week prior to travel, continued throughout the stay in the malarious region and taken for 4 weeks after returning home (British National Formulary, 1998).

The feasibility of conducting a prospective study of travellers to The Gambia had been established by Clift et al. (1997) who surveyed 113 UK tourists. Only four respondents reported not taking any chemoprophylaxis and amongst those who were, three-quarters were taking mefloquine and just over a fifth were taking a combination of chloroquine and proguanil. 78 tourists were successfully followed up and adherence rates were found to be significantly higher among those taking mefloquine (82%) than for those taking chloroquine/proguanil and other preparations (38%).