This study provided an opportunity to assess the influence of health workers’ attitude to the usage of new malaria treatment recommendations. Overall, results showed variations in health workers attitudes and practices regarding new treatment recommendations in terms of type of health facility, ownership and type of health worker at six months post changes and two years later. There was less variation of provider’s attitudes and personal use of new recommended antimalarial between districts. Dispensaries and health centers showed higher preferences for SP than hospitals. Similarly, public facilities reported higher preference for SP than non-public ones. Most providers were not comfortable with the use of SP for children below age of 5 and for pregnant women. Personal use of SP for their malaria episodes was high in both rounds, with some exceptions in the second round.
The introduction of SP as a first line treatment of uncomplicated malaria in Tanzania was not well received
. It is therefore not surprising that from the first survey (baseline), preference to SP as appropriate treatment for management of uncomplicated malaria was low among health workers in the surveyed districts. Their dissatisfaction of treatment recommendation could have influenced their perceptions, attitudes and practices. This attitude may be based on their daily experiences in the clinical management of patients as became evident when we assessed providers’ preferences by type of health facility. In many areas especially rural Tanzania, hospitals are expected to be receiving referal cases. For malaria, these could mean patients who did not respond well to first line treatment at lower level of care, i.e. dispensary or health center, hence sent to hospitals for further management including laboratory assessment and in-patients’ service. This may explain the significant findings of providers from dispensaries and health workers appreciating SP use better than hospitals.
Many studies have assessed users’ perceptions of new treatments when changes occur. Several authors explored community perceptions to malaria treatment and other aspects of health services in Tanzania and elsewhere
[15–17]. Likewise, most studies of health workers’ knowledge, perceptions, attitudes and understanding have been conducted in relation to health services and health problems other than malaria
[18–20]. The introduction of ACTs in most African countries received considerable attention, with researchers evaluating the process of change and performance of health workers on new policies. Some of these evaluations were on artemether-lumefantrine in Kenya, Uganda, and Zambia as well as on artesunate plus amodiaquine in Ghana
[14, 21–23]. These evaluation assessed providers’ use of new treatment recommendations for malaria case management, with no focus on personal preferences and personal use.
The difference in providers’ preferences for SP for management of uncomplicated malaria may also be related to performance of the health facilities governing committees. A fact that public providers were more comfortable with SP than those in the non-public sector may be linked to a closer supervision of health management teams. Intrinsically health workers do assess clinical progress of their patients. Results in this survey indicate that most providers were skeptical using SP for children under 5, and pregnant women; probably because they perceived it too strong for children below age of 5 as a previous study from Tanzania reported
.This preference worsened over time suggesting that providers were not satisfied with experiences of using SP. This finding is in contrary to what one would expect; that providers need time to appreciate, accept and comply with new treatment policies. An important lesson here is that, when there is a failing drug in the system, health care providers will, without doubt, notice it and may provide initial indication of the drug resistance in the population.
Despite poor attitudes to the new drug for first line management of uncomplicated malaria, many providers indicated that they had used SP in their last illness episode of malaria. In the first round, it was difficult to assess if SP was used before or after the change, since we did not specify the duration of illness prior the survey, but in the second round we gave a time frame; i.e. we inquired for a malaria episode in the past three months preceding the survey. Also, we did not seek additional clinical information; therefore couldn’t assess if it was correctly used. Interestingly, compared to hospitals, providers’ from the dispensaries were more likely to have used SP for their illness episode or their family members. This finding was observed in both survey rounds. One possible explanation for this observation may be related to a fact that, hospitals are a higher level of care, therefore more likely to see referal cases of malaria; i.e. non response to first line treatment and/or severe form of the disease. But also, dispensaries do not have a wider range of treatment choices and services available, hence more likely to follow treatment guidelines presented.
Also, it is more likely that, knowing this is the only available treatment option for them, dispensaries strives to have medications available in stock; hence availability of the drugs facilitated it being used by a staff or staff’s family member. This may not be a case for higher levels of care, given a wider choice of drugs available. The same may apply for public providers, with good health management team supervision, public facilities are more likely to abide by the new treatment recommendations, but this cannot be said for non-public facilities, hence significantly less use of SP for last malaria episodes was observed from non-public providers in this study. SP was available as a single dose and its price was not as high as other antimalarials available at the time of survey. During this time, other antimalarials available in private sector included amodiaquine, chloroquine (the outgoing medicine), artemisinin mono-therapy, quinine, etc. All of these products require more than a single dose to finish a course of treatment, therefore more likely to cost more than SP. A possibility of financial gain for using other treatment recommendations than SP cannot be ruled-out in the private sector.
We did not account for the clustering of health facilities in the analysis. This may have affected the magnitude of the measured effect. Ideally, this clustering effect should be taken into account because there may be similarities between individuals working in the same health facility, such that, on average they are more similar to each other than to individuals in other health facilities, due to many factors such as training received at facility level and experiences acquired through everyday’s practices. However, we worked with the assumption of independence between the observations, since we were assessing individuals’ attitudes, through their preferences and personal use of treatment recommendations. These variables are more likely to be related to personal understanding and beliefs.
However, it is acknowledged that personal preferences can be influenced by many factors such as training, work experience and for the case of malaria treatment; availability of medicines and appropriate technologies to assist in clinical care of patients e.g. diagnostics for malaria confirmation, as well as presence of policy briefs and documents for referencing. These factors were not assessed and therefore limit our conclusions with regards to the role they play to shape health workers preferences and personal use of new treatment recommendations for management of uncomplicated malaria in the surveyed area.
Third, a fact that the criteria used to obtain interviewee was not random, implies that results from this evaluation cannot be generalized for all health workers in Tanzania. However, we are confident that, this study provided additional information on predictors of preferences and practices among health care providers toward SP, which complimented previous reports of poor community and provider’s perceptions towards SP when it was introduced for management of uncomplicated malaria in Tanzania; as well, it provides a clue on what happens to the health system when there is a failing drug.
Although we did not match respondents in the two surveys, the fact that we interviewed health workers from the same health facilities, increased our confidence that the differences reported in preferences and practices reflect a general picture for providers with similar experiences.
Fourth, not being able to assess clinical information when assessing practices towards recommended treatment through personal/family use, might have led to a biased estimation of SP use. It is possible that there were good reasons for not using SP to some cases that may be due to, for example, a diagnosis of severe malaria, non-response to SP or history of hypersensitivity reaction to sulphur- containing medicines.
Fifth, it is possible that some health workers reported what was considered appropriate rather than what they would actually do, or actually did, leading to courtesy bias. Furthermore, recalling what happened in terms of treating malaria in the past may have been difficult for some participants, introducing a recall bias. These biases could have affected measures of effect estimated. In this respect, we limited the recall for up to the past three months in the second survey.
Lastly, the relatively small sample for some sub-groups of explanatory variables e.g. physicians; made it difficult to detect associations between some potential risk factors and the outcomes studied in those groups.