The authors have declared that no competing interests exist.
Conceived and designed the experiments: JJ CVG. Performed the experiments: YK BB JPA DN HD. Analyzed the data: KT. Wrote the paper: KT JJ.
Antibiotic resistance (ABR) particularly hits resource poor countries, and is fuelled by irrational antibiotic (AB) prescribing. We surveyed knowledge, attitudes and practices of AB prescribing among medical students and doctors in Kisangani, DR Congo.
Self-administered questionnaires.
A total of 184 questionnaires were completed (response rate 94.4%). Knowledge about AB was low (mean score 4.9/8 points), as was the estimation of local resistance rates of
This data shows the need for interventions that support rational AB prescribing.
Increasing resistance of bacterial pathogens to commonly used antibiotics (AB) has become a world-wide public health concern
The causes that drive ABR in low income countries are well known. The main factors are irrational drug use such as over-prescription and unnecessary prescription of AB (such as for viral infections), incomplete treatments and self-medication as well as insufficient infection control measures to prevent spread of resistant bacteria both in the community and the hospital
There is only little KAP-information among prescribers in hospital settings
The survey was conducted during March – April 2011 in the area of Kisangani, a town of 600,000 inhabitants in the Oriental Province, in the center of Democratic Republic of the Congo (DRC) (
The University Hospital of Kisangani (UH Kisangani) is a 129-bed referral and teaching hospital accommodating students of the Faculty of Medicine of the University of Kisangani for clinical clerkships. Despite on top of the health pyramid, UH Kisangani is only basically equipped as a consequence of recent wars and a degraded economic situation. There are no functioning microbiology services and no antibiotic committee. Bed occupancy rate in 2011 was estimated at 45.5% (UK Kisangani, annual report 2011). The other health centers and the two referral hospitals in Kisangani have similar constraints of infrastructure and equipment.
This study was performed as a cross-sectional survey with purposeful sampling among last year medical students (who are prescribing as part of their practice) as well as prescribing medical doctors with different extent of work experience. The survey was a self-administered questionnaire which was distributed and collected by trained collaborators (co-authors Y.K. and B.B.).
Last year medical students were recruited at UH Kisangani (total number of students was 131). Prescribing medical doctors were visited on-site at the hospitals or health centers after having made an appointment. Medical doctors working in the administrative services were not addressed. According to the Provincial Health Division, the total number of prescribing medical doctors in Kisangani in 2011 was estimated at 185 (personal communication).
The surveys were given to the respondents who filled in the forms immediately. During completion, the collaborator was available for explanation. The filled-in copies were handed over immediately after completion. No incentives were given to the respondents.
The survey instrument was developed according to a previous questionnaire used in Peru
The original questionnaire was amended with questions addressing the resources about AB prescribing (including pharmaceutical companies), the perceptions of the causes of ABR and the quality of locally available AB (one question each). The final survey consisted of 33 questions (
Pending the installation of an ethical committee in the Oriental Province and at the UH Kisangani, the study was assessed and approved by the Provincial Health Officer and the Director’s Board of UH Kisangani, the highest possible authority. All data was analyzed anonymously.
Data were anonymously entered in an Excel database (Microsoft Corporation, Redmond, Washington, USA). Proportions and means were calculated for categorical and continuous variables respectively. Chi square, Fisher's exact test and Mann-Whitney U test were performed as appropriate to calculate a 95% level of significance. All statistical analyses was performed using STATA version 12 (Statacorp Texas, USA).
Eighty-nine of the estimated 185 medical doctors in Kisangani were addressed, as well 106 of the 131 medical students at UH Kisangani (coverage rates of 48.1% and 80.9% respectively). Of the total of 195 questionnaires, 11 (5.6%) were either not returned or incompletely filled in (all from medical doctors), leaving 184 eligible surveys (response rate 94.4%). Medical doctors and students accounted for 78 (42.3%) and 106 (57.6%) of respondents respectively.
Among the medical doctors, 11.5% (n = 9) had less than one year work experience, 38.5% (n = 30) had between 1 and 3 year work experience, 23.1% (n = 18) between 4 to 6 years work experience and the remaining 26.9% (n = 21) had more than 10 years of work record. Unless otherwise stated, the results are presented for all respondents (medical students and medical doctors regardless of work experience) combined, except when there were differences between these groups.
All respondents prescribed AB to in- as well as outpatients. More than half (100/183, 54.4% - one respondent did not reply to this question) reported to prescribe AB more than once a day, 15.2% (n = 28) and 17.4% (n = 32) reported 1 to 2 times and 3 to 5 times per week respectively. As expected, medical students prescribed less frequently compared to medical doctors (≥1 prescription/day: 42.5% versus 85.7% and <1 prescription/day: 57.5% versus 14.3%; p<.0001).
On a score of 8, the mean (± S.D.) and median (range) scores were 4.9 (±0.09) and 5 (2–8) respectively. A total of 1 (0.5%) and 20 (10.9%) of respondents scored 8 and 7 points respectively; scores of 6 and 5 were achieved by 36 (19.6%) and 55 (29.9%) respondents respectively. The remaining respondents (n = 72, 39.1%) had a score of 4 or less. There was no significant difference in the mean score according to years of experience (p = .19), frequency of consulting colleagues before deciding about AB prescribing (p = .21) or previous training about AB prescribing (p = .54).
As to the individual questions (
No. | Question and |
Overall N (%) answered correctly |
N (%) of doctors answered correctly |
N (%) of students answered correctly |
p |
1 | A 4-year-old girl has diarrhea for 4 days (3 stools/day). She has no fever at examination nor duringthe last few days. Which treatment do you propose? | ||||
|
165/184 | 68/78 | 97/106 | 0.340 | |
|
(89.7%) | (85.9%) | (74.5%) | ||
|
|||||
|
|||||
2 | A 6-year-old child has a fever of 38°C, purulent rhinitis and angina for two days. At inspection,the throat is reddish. Which treatment do you recommend? | ||||
51/183 | 28/78 | 23/105 | 0.038 | ||
|
(27.9%) | (35.9%) | (21.9%) | ||
|
|||||
|
|||||
3 | During your ward round, you see two patients with severe renal failure. Patient A is a 68-year-old man suffering from serious cellulitis at the leg, he is treated with clindamycin. Patient B is a 64-year-olddiabetic woman which is blindly (empirically) treated for septicemia with ceftriaxone. Dosage reductionis needed for: | ||||
|
34/183 | 12/77 | 22/106 | 0.375 | |
|
(18.6%) | (15.6%) | (20.7%) | ||
|
|||||
|
|||||
4 | Which one of the following antibiotics is safe during pregnancy? | ||||
|
170/183 | 76/78 | 94/105 | 0.039 | |
|
(92.9%) | (97.4%) | (89.5%) | ||
|
|||||
5 | Which one of the following antibiotics has the best activity against anaerobes? | ||||
|
171/182 | 72/77 | 99/105 | 0.827 | |
|
(94.0%) | (93.5%) | (94.3%) | ||
|
|||||
6 | Methicillin resistant |
||||
|
64/174 | 33/73 | 31/101 | 0.0500 | |
|
(36.8%) | (45.2%) | (30.7%) | ||
|
|||||
|
|||||
7 | Which one of the following antibiotic most effectively crosses the blood-brain barrier? | ||||
|
132/179 | 58/75 | 74/104 | 0.354 | |
|
(73.7%) | (77.3%) | (71.1%) | ||
|
|||||
8 | Aminoglycosides such as gentamicin are very active if they are administered as follows : | ||||
|
109/179 | 44/74 | 65/105 | 0.741 | |
|
(60.9%) | (59.5%) | (61.9%) | ||
|
|||||
9 | In DR Congo, what is according to your information the estimated resistance rate of |
||||
|
76/179 | 32/75 | 44/104 | 0.962 | |
|
(42.5%) | (42.7%) | (42.3%) | ||
|
|||||
|
|||||
10 | In DR Congo, what is according to your information the estimated resistance rate of |
||||
|
12/173 | 3/70 | 9/103 | 0.258 | |
|
(6.9%) | (4.3%) | (8.7%) | ||
|
|||||
|
only questions 1–8 were included in the score;
trimethoprim/sulphamethoxazole.
Less than half and fewer than 10% of respondents correctly estimated the local resistance rate of,
The majority of respondents agreed or strongly agreed that ABR is an important problem, though significantly (p<.0001) less in their own practice as compared to the national or worldwide scope (
Antibiotic resistance is a problem | World wide | In DRC | In my practice |
n (%) | n (%) | n (%) | |
strongly agree | 91 (49.5%) | 115 (62.5%) | 55 (29.9%) |
agree | 66 (35.9%) | 56 (30.4%) | 69 (37.5%) |
neutral | 16 (8.7%) | 8 (4.4%) | 20 (10.9%) |
do not agree | 4 (2.2%) | 2 (1.1%) | 26 (14.1%) |
do not at all agree | 6 (3.3%) | 3 (1.6%) | 13 (7.1%) |
no answer | 1 (0.5%) | 0 (0%) | 1 (0.5%) |
Treatment not completed | Antibiotic not adapted | Too low dosage | Poor quality AB | Too much prescription and consumption | In-hospital transmission | Self-medication | |
Yes | 89.7% | 83.2% | 82.1% | 71.2% | 69.0% | 64.1% | 94.6% |
DK |
3.8% | 6.0% | 8.2% | 13.6% | 11.4% | 12.5% | 3.3% |
No | 4.4% | 9.8% | 8.2% | 10.9% | 16.9% | 17.9% | 1.6% |
ND |
2.2% | 1.1% | 1.6% | 4.4% | 2.7% | 5.4% | 0.5% |
Don’t Know.
No Data.
Nearly 90% of the respondents declared to feel very confident (n = 21, 11.4%) or confident (n = 142, 77.2%) about their knowledge on AB. Medical students tended to have lower self-confidence in AB prescribing than medical doctors (85.7% versus 94.8%), although this difference did not reach statistical difference (p = .053).There was no significant correlation between confidence and mean knowledge score (p = .1622). Despite this confidence, more than half (55.4%) of respondents agreed with the statement that the selection of the correct AB is difficult. However, when asked about the frequency of consulting a colleague when prescribing AB, most respondents replied sometimes (79.4%) and never (1.1%) versus 3.3%, 11.4% and 4.0% replying half of the times, mostly and always respectively. In line with the lower self confidence among student they also reported to consult significantly more often a colleague compared to medical doctors (25.4% versus 11.6%; p = .019).
For medical doctors and students combined, the sources of information about AB used during the month prior to the survey were, in rank of decreasing frequency: pharmaceutical companies (consulted by 73.9% of respondents), antibiotic guidelines (66.3%), university courses (63.6%) and information retrieved from the internet (45.7%). Only 26.6% of respondents declared to have used WHO guidelines.
Source of information | Numbers (%) of medical students (n = 106) | Numbers (%) of medical doctors (n = 78) | p |
Pharmaceutical companies | 76 (71.7%) | 60 (76.9%) | 0.425 |
University courses | 88 (83.0%) | 29 (37.2%) | <.0001 |
Internet | 44 (41.5%) | 40 (51.3%) | 0.188 |
Antibiotic guidelines | 73 (68.9%) | 49 (62.8%) | 0.391 |
WHO guidelines | 29 (27.3%) | 20 (25.6%) | 0.795 |
Patient pressure was perceived as a factor contributing to overuse of AB in the community by nearly two-thirds (61.9%) of respondents, whereas only one third (34.3%) did so for the hospital setting (p<.001). The majority of respondents (n = 165; 89.7%) agreed that knowledge of local AB resistance patterns was needed for good prescribing. About two-thirds (63.8%) of respondents did not agree with the statement that choice of prescribed AB was more influenced by availability than by the cause of infection. A total of 72.3% of respondents disagreed with the statement that ABs if not needed do not cause harm to the patient; however, 6.0% were neutral and about one fifth (21.7%) subscribed this statement.
Only 16.3% (n = 30) of respondents agreed that ABs available at the local pharmacies were in most cases of good quality (39.7% mentioned they did not have an opinion about it). One third (30.4%) believed that AB coming through the national central procurement office were of good quality (i.e. significantly higher than those available at the pharmacies; p<.001), whereas 22.8% considered them to be of poor quality and nearly half (44.6%) of respondents did not indicate an opinion.
Nearly three quarters (73.4%) of respondents agreed with the statement that local AB guidelines were more useful than international guidelines. AB committees were only seen by 9.8% of respondents as obstacles rather than as a help, although 42.9% of respondents declared not to have an opinion about this topic. Nearly all (98.8%) respondents expressed their wish for additional training on AB prescribing.
KAP-surveys provide insights in the driving forces of AB prescribing and are complementary to surveillance of AB resistance and AB consumption which together constitute baseline information to design interventions about AB prescribing
To counter these limitations, the present survey was validated by a team of local professionals and took into account limitation in time and concentration of the respondents. In order to avoid influence on answers regarding attitudes and practice questions assessing knowledge were presented at the end of the questionnaire. In order to minimize professional pressure, students were individually addressed by one of their peers (Y.K.). In addition, the questionnaire was distributed and filled-in on-site, precluding consultancy of peers or resources.
Among the strengths of the present KAP-survey was the fact that it addressed most of the topics cited by WHO as influencing AB prescribing, except for economic incentives and workload
The overall mean score on the knowledge questions (4.9/8 points) was lower than the result from the similar survey conducted in Peru
Further, there was insufficient knowledge about local AB resistance rates of
Similar to previous surveys
Despite the fact that more than half of respondents agreed that it is difficult to choose the correct AB, there was high self-confidence in AB prescribing which seemingly contrasted to the modest knowledge scores. Likewise, there was a low propensity to consult colleagues when prescribing AB. Medical students tended to be less self-confident and were more willing to seek input of colleagues. Similar difference in attitudes have been observed between residents (doctors in training) and medical doctors and these observations confirm medical students and junior doctors as a feasible target group for future interventions in AB prescribing
The availability of unbiased information about AB is a prerequisite for appropriate AB prescribing
Patient pressure was reported as major factor for AB prescription in the DRC although significantly less in the hospital compared to the community setting. In the Peru survey, patient’s demand was even stronger perceived as a thriving factor with more than 70% in the community setting and 50% in the hospital setting.
The fact that local AB resistance patterns and the cause of infection rather than the availability of AB were scored as important to the majority of respondents is promising in regard to future interventions. Most of concern however was that more than a quarter of respondents did not contest the statement that unnecessary AB are not of harm to the patient.
AB coming through the national central procurement office
In line with the demand for local AB resistance rates, nearly all respondents welcomed local AB guidelines. AB committees are not yet functioning in DR Congo and this was apparent from the absence of opinions about this topic.
The present study shed a light on the AB prescribing behavior by medical doctors and students in resource-limited settings in Central Africa. In order to understand the whole extent of inappropriate AB prescribing and usage, further research is needed among the other channels of AB prescribing and procurement, as well as from private dispensers and the general public. Differences between the results of previous surveys in other countries
(DOC)
We want to thank Dr. Coralith Garcia, Institute of Tropical Medicine Alexander von Humboldt, Lima Peru for sharing with us the questionnaire used in reference 9.