bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
1 Health seeking behaviour and cost of fever treatment to households in a malaria endemic
2 setting of northern Ghana: A cross sectional study
3
4 Maxwell Ayindenaba Dalaba1*, Patricia Akweongo2, Philip Baba Adongo2, Philip Ayizem
5 Dalinjong1, Samuel Chatio1, and Abraham Oduro1
6
7
8 1 Navrongo Health Research Centre, P.O Box 114, Navrongo, Ghana
9 2 University of Ghana, School of Public Health, Accra, Ghana
10
11
12
13
14
15
16 Corresponding Author
17 Maxwell Ayindenaba Dalaba, Navrongo Health Research Centre, P.O Box 114, Navrongo,
18 Ghana
19 Email: [email protected]
20
21
22
23
1 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
24 Abstract
25 Background: This study examined the health seeking behaviours and cost of treatment of malaria
26 to households in the Kassena-Nankana district of Ghana.
27 Methods: A cross-sectional household survey was conducted between July and September 2015.
28 Individuals who had an episode of fever or malaria in the past two weeks identified during routine
29 Health and Demographic Surveillance System data collection were selected for the study. Socio-
30 demographic characteristics, treatment seeking behaviours and cost of treatment of malaria were
31 obtained from the patient perspective.
32 Results: Out of the 1,845 households visited, 21.3% (393/1,845) reported to have had an episode
33 of fever or malaria in the past two weeks. Of the 393 people with malaria, 66.9% (263/393)
34 reported taking an antimalarial. About 53.6% (141/263) of the antimalarials were obtained from
35 formal healthy facilities. About 36.1% (95/263) reported to have taken Dihydroartemisinin-
36 piperaquine, 35.4 % (93/263) took Artesunate–Amodiaquine and 21.7% (57/263) took
37 Artemether-Lumefantrine. Only 49.6%% (195/393) of the study participants had their blood
38 sample taken for illness (microscopic or Rapid Diagnostic Test). Only 23.6% (62/263) took
39 antimalarial within 24 hours of the onset of illness. The overall average costs (direct and indirect
40 cost) incurred by households per malaria treatment was GH¢27.82/US$7.32 (range: GH¢0.2/
41 US$0.05 - GH¢200/ US$52.63). The average cost incurred in the treatment of malaria by those
42 who were enrolled into the National Health Insurance Scheme (NHIS) was GH¢24.75/US$6.51
43 and those not enrolled was GH¢43.95/US$11.57.
44 Conclusions: Prompt treatment such as treatment within 24 hours of onset of malaria was low.
45 The average costs to households per malaria treatment was GH¢27.82/US$7.32 and the preferred
2 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
46 antimalarial was Dihydroartemisinin-piperaquine. There was a positive effect of NHIS enrolment
47 on cost of treatment as the insured incurred less cost (US$5 less) in treatment than the uninsured.
48 Keywords: Malaria, fever, treatment-seeking behaviour, cost, Ghana.
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
3 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
69 Introduction
70 According to the WHO report, there are an estimated 198 million malaria cases and 584 000
71 malaria deaths worldwide in 2013. The burden of malaria is more severe in sub-Saharan Africa
72 where about 90% of all malaria deaths occur [1]. It is a major cause of poverty and slows economic
73 growth by up to 1·3% per year in endemic countries[2].
74 Malaria is a significant public health problem in Ghana where it accounts for about 33% of all
75 outpatient attendances and 49% of under five years admissions[3]. Malaria is a preventable,
76 diagnosable and treatable illness. The main methods for the prevention of malaria include Long-
77 Lasting Insecticidal Nets (LLINs), Indoor Residual Spraying (IRS), Intermittent Preventive
78 Treatment for pregnant women (IPTp) and vector control methods such as larviciding [1,4]. The
79 WHO has recommended that all persons from endemic areas with suspected malaria should be
80 examined for evidence of infections with malaria parasites by Rapid Diagnostic Test (RDT) or
81 microscopy. In addition, WHO recommended that uncomplicated malaria should be treated with
82 antimalarial such as Artemisinin-based combination therapies (ACTs), particularly in areas where
83 malaria is endemic and persons should have access to ACTs within 24 hours of onset of malaria
84 [1].The recommended ACT combinations are artemether-lumefantrine (AL), artesunate-
85 amodiaquine (AS+AQ), artesunate-mefloquine (AS+MQ), dihydroartemisinin-piperaquine (DP),
86 and artesunate-sulfadoxine-pyrimethamine (AS+SP). The choice of ACT in a country or region is
87 based on the level of resistance of the partner medicine in the combination [5,6].
88 Following WHO recommendation, in 2004 Ghana changed her antimalarial drug policy choosing
89 Artesunate-Amodiaquine combination as the first line drug for the treatment of uncomplicated
90 malaria to replace chloroquine. However, the implementation process was confronted with some
91 challenges which relates to adverse drug reactions, safety concerns and absence of other treatment
4 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
92 options. It therefore became necessary to review the new policy to address all recognized concerns.
93 A task force reviewed the existing policy and selected additional ACT drugs and dosage forms to
94 accommodate for those who could not bear the Artesunate- Amodiaquine combination. Two
95 additional first line ACTs namely; Artemether-Lumefantrine and Dihydroartemisinin-Piperaquine
96 were selected thus making Ghana to have three official concurrent first line antimalarial treatments
97 [7].
98 To date, however, the reaction of the health system to multiple alternative first line malaria therapy
99 has not been comprehensively studied. Little is known about access to effective and prompt
100 treatment and preferences for antimalarial drugs for treatment of fever/malaria as well as cost of
101 malaria treatment in the face of multiple first line treatment policy. This study therefore examined
102 the health seeking behaviours as well cost of treatment of malaria to households in the Kassena-
103 Nankana district of Ghana in the face of multiple first line antimalarial drugs.
104
105 Methods
106 Study area
107 The study was carried out in the Kassena-Nankana East and West Districts of the Upper East
108 Region of Ghana. The Kassena-Nankana East is home to the Navrongo Health Research Centre
109 (NHRC) and staff of the Centre conducted the study. The NHRC operates Health and Demographic
110 Surveillance System (HDSS) and has a data base of all individuals and households in the Kassena-
111 Nankana East and West Districts. For the purpose of research, the NHRC refers to the two districts
112 by their former name - the Kassena-Nankana District (KND).
5 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
113 The KND covers an area of about 1,674 square kilometres of land [8]. The area is characterized
114 by a short rainy season and a prolonged dry season from October to March. The mean annual
115 rainfall is about 850mm, with the heaviest usually occurring in August. The total population
116 currently under surveillance is about 152 000 residing in about 32,000 households [8]. Malaria
117 transmission in the KND occurs all year round but there is a distinct seasonal pattern with the peak
118 of transmission coinciding with the period of the major rains (August) and the dry season seeing
119 very low rates of malaria infection[9].
120 The districts have one district referral hospital located in Navrongo town and 8 health centres
121 strategically located across the district which provide secondary curative and preventive health
122 care. There are 28 Community Based Health Planning and Services (CHPS) compounds/clinics
123 located in various communities and providing primary health care treatment for minor ailments
124 and also carrying out childhood immunizations and antenatal services. There are three private
125 clinics, three pharmacies and over 50 licenced chemical shops in the area.
126 Study design
127 A cross-sectional household survey design was used, and data was collected using quantitative
128 approach. The study was conducted within the INDEPTH-network effectiveness and safety studies
129 (INESS) Phase IV research platform that aimed to assess the effectiveness of new malaria
130 treatments and its determinants in real life health systems[10,11].
131
132 Sample size and data collection
133 A standardized, rolling, nested, household survey was used, by employing two-week recall period.
134 We assumed conservatively that at least 6% of the population would have had fever event two
6 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
135 weeks prior to the interview date and that about 50% of the population with fever gained access to
136 an official point of ACT provision within 24 hours. We accounted for clustering within households
137 by using a design effect of 1.5. We further allowed a 10% drop out rate to achieve a sample size
138 of 1, 845 households.
139
140 The data for this study were collected as an additional module within the framework of the
141 Navrongo Health and Demographic Surveillance System round updates. Field workers visited the
142 sampled households between July and September 2015. All members in the selected households
143 with a history of fever in the prior two weeks were interviewed using a structured questionnaire.
144 For children who were less than 18 years of age their adult caretakers were interviewed.
145 Information on socio-demographic characteristics, treatment seeking behaviours and cost of
146 treatment were obtained. Cost of treatment was collected from the patient perspective.
147 Those who treated with antimalarial drugs were asked the name of the drug and a physical
148 observation of the drugs if it was still in use or if the package was kept. Field workers carried photo
149 images of the variety of antimalarial packages (brand and generic names) available from private
150 and public providers in the district to facilitate identification of treatments received. Types of
151 antimalarial drugs were to be captured on the questionnaire by their generic names. For instance,
152 brands such as Camosunate, Amonate, Coarsucam, to be captured as Artesunate-Amodiaquine;
153 Malar-2, Lonart , lumartem, Coartem, Co-Malagon, Lumether as Artermether-Lumefantrine; and,
154 P-Alaxin, Eurartesin, Duo-Cotecxin, as Dihydroartemisinin-piperaquine.
155
156
7 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
157 Data processing and analysis
158 Data were double entered into fox pro and transferred to Stata 11.0© for analysis. We used
159 descriptive statistics showing proportions, means and numbers.
160
161 Ages of respondents were captured as continuous variables but during analysis were grouped into:
162 under 5 years, 5-14 years (representing children), 15-49 years (representing reproductive age), and
163 49 years and above.
164 Insurance status of the patient was defined as having a valid national health insurance card on the
165 day of interview. The types of health care providers that patients sought care from were grouped
166 into three categories: formal health facilities (comprising hospital, clinics, health centres and
167 CHPS), pharmacy shops/licenced chemical shops, and home-based care (left over drugs and
168 traditional care).
169
170 The wealth of the household was measured in terms of household assets and possessions using
171 principle component analysis (PCA). We did not collect data on household assets and possessions
172 in our study, but rather used data collected by the Navrongo Health and Demographic Health
173 Surveillance System (NHDSS). The NHDSS database has information on all households in the
174 study district[8]. The study respondent’s perm IDs were linked to the NHDSS data to generate the
175 household wealth index using PCA technique.
176
177 A total of 53 households could not be linked with the Navrongo HDSS household assets database
178 and therefore could not be classified and included in the analysis. Thus, household wealth of 340
179 households were generated. The assets and possessions used for the construction of the wealth
8 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
180 index comprised of the type of material used for the wall of the household, roofing material,
181 cooking utensils, toilet facility, source of drinking water and cooking fuel. Household possessions
182 included bicycle, motorbike, car, radio, bed, sewing machine, tape player, television, Digital
183 Versatile Disc (DVD), mobile phone, refrigerator, cattle, sheep, goat, pig and donkey. Households
184 were then assigned to five quintiles: least poor, less poor, poor, very poor, and poorest.
185
186 The total cost of fever/malaria per household was calculated by summing the direct and indirect
187 costs incurred by households. The direct cost included both direct medical cost and non-medical
188 costs. The direct medical costs covered all out-of-pocket payments (OOP) for registration card,
189 consultation, diagnosis, medicines and medical supplies on the patient; and direct non-medical
190 costs included all out-of-pocket payments for transportation to and from health facilities (patient
191 and caregiver) and special foods for patients.
192
193 Indirect cost (productivity lost) was estimated by multiplying the number of days lost due to
194 malaria by the daily wage for year 2015 (GH¢6/US$1.6). This was calculated for both the patient
195 and the caretaker who were actively engaged in an income-generating activity prior to the illness.
196 In-patient or admission was defined as those who were detained in a health facility for malaria for
197 longer than 12 hours. In-patient costs were therefore the sum of medical costs, laboratory costs,
198 and bed costs during the admission. All costs were collected in Ghana cedis (GH¢) and results
199 presented in both Ghana cedis and US$. The conversion of Ghana cedis to US$ was based on the
200 average exchange rate for 2015 at US$1=GH¢3.8.
201
202
9 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
203 Ethical consideration
204 Ethical approval for the study was obtained from the Navrongo Health Research Centre
205 Institutional Review Board and the Ghana Health Service Ethics Committee. Written informed
206 consent was obtained from all adult participants and caregivers/parents of children before
207 interviews were conducted. In addition, assent was sought from participants who were between
208 12-17 years.
209
210 Results
211 Background characteristics
212 Table 1 shows the background characteristics of respondents. Out of the 1,845 households visited,
213 21.3% (393/1,845) reported to have had an episode of fever or malaria in the past two weeks. The
214 average age of fever patients was 25years and majority of the patients 45% (177/393) were above
215 49 years old and 26% were under-five years of age. Majority of the fever cases were females 51.7%
216 (203/393) compared to 48.4% (190/393) males. About 83.2% (327/393) of patients had valid
217 national health insurance scheme (NHIS) cards at the time of illness. About 32.3% (127/393) of
218 patients were actively engaged in income generation activity. Out of those who were actively
219 engaged in income generation activity, majority 40.2% (51/127) were farmers, 37.8% (48/127)
220 were artisans, and 22.1% (28/127) were public servants. Most patients 42.8% (168/393) were from
221 poor households (lowest two quintiles).
222
223
224
225
10 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
226 Table 1: Background characteristics of patients who reported fever/Malaria Characteristics Categories Number Percent (%) (N=393) Under-five 102 26.0 5 -14 years 65 16.5 Age 15-49 49 12.47 Above 49 177 45.02 Male 190 48.35 Sex Female 203 51.65 Insured Insurance status 327 83.21 Uninsured 66 16.79 Actively engaged in Yes 127 32.32 income generating No 266 67.68 activity Occupation of those Farmer 51 40.15 actively engaged in Public service 28 22.05 income generating Artisan activity 48 37.8 Poorest 82 20.9 Very poor 86 21.9 Poor/middle 46 11.7 Household wealth Less poor 52 13.2 Least poor 74 18.8 Missing 53 13.5 227 228 229
230 Malaria diagnoses and treatment seeking behaviours
231 Out of the 393 fever patients interviewed, 66.9% (263/393) reported taking antimalarial (ACT).
232 Most of those who used antimalarials 53.6% (141/263) obtained it from formal health facilities
233 (hospitals, Health centres, clinics and CHPS), 41.4 % (109/263) from pharmacy or chemical shops
234 and 5% (13/263) at home. With regards to the type of antimalarial patients used, 36.1% (95/263)
235 reported to have taken Dihydroartemisinin-piperaquine, 35.4% (93/263) took Artesunate–
11 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
236 Amodiaquine, 21.7% (57/263) took Artemether-Lumefantrine. Only 49% (195/393) of the study
237 participants had their blood sample taken for illness (microscopic or RDT) and subsequently
238 73.3% (143/195) were confirmed to have malaria (Table 2).
239 Table 2: Malaria diagnoses and treatment Variables Categories Frequency Percent (n=393) (%) Antimalarial 263 66.9 Type of treatment Other treatments 130 33.1 Yes 195 49.2 Blood sample taken No 198 50.8 Yes 143 73.3 Confirmed malaria No 52 26.7 Health facility 141 53.6 Place where ACT Licenced chemical shop 109 41.4 was obtained Home 13 5 Artesunate–Amodiaquine 93 35.4 Artemether-Lumefantrine 57 21.7 Type of Antimalarial used Dihydroartemisinin-Piperaquine 95 36.1 Other Antimalarials (single dose) 18 6.8 240
241 With regards to prompt treatment, majority of patients did not take antimalarial within 24 hours.
242 For instance, among those who took antimalarial, only 23.6% (62/263) took antimalarial within 24
243 hours, whiles 44.5% (117/263) took antimalarial between 25 to 48 hours and 31.9% (84/117) took
244 antimalarial after 48 hours of onset of malaria (Table 3).
245
246
247
248
12 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
249 Table 3: Time spent between onset of fever and start of antimalarial treatment 250 Time spent before start of antimalarial Frequency (n) Percentage (%) Within 24 hrs 62 23.6 Between 25 to 48 hrs 117 44.5 more than 48 hrs 84 31.9 Total 263 100.0 251 252 253
254 Cost of treating malaria
255 Table 4 presents direct and indirect costs of treating malaria. Out of the 393 patients, 25.5%
256 (140/393) patients incurred direct medical costs during malaria treatment. The cost was calculated
257 for only those who incurred cost (OOP)in treating the malaria. The average direct medical cost
258 was GH¢11.91 (US$3.13) per treatment. Out of the 140 patients that incurred direct medical costs,
259 74.3% (104/140) were insured and 25.7% (36/140) were not insured. The average direct medical
260 cost by the uninsured was higher (GH¢14.91/US$3.92) than the insured (GH¢10.87/US$2.86).
261 Out of the total of 393 patients, 82.2% (323/393) spent on direct non-medical costs during the
262 malaria treatment. The direct non-medical coats included transportation and other costs such as
263 special food, water and lodging. About 77.4% (304/393) of respondents incurred transportation
264 cost. The average transportation cost to and from health care providers was estimated at GH¢10.40
265 (US$2.74). The average direct non-medical costs (transportation and other costs such as special
266 food, water and lodging) was GH¢11.69 (US$3.08). The average direct cost representing medical
267 and non-medical cost was calculated as GH¢14.87/US$3.91 (range: GH¢0.2/US$0.05 -
268 GH¢166/US$43.68). Average indirect cost which relates to productivity lost was estimated at
13 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
269 GH¢25.04 (US$6.59). The overall average costs (direct and indirect cost) incurred by households
270 per malaria treatment was GH¢27.82/US$7.32 (range GH¢0.2/US$0.05 - GH¢200/US$52.63).
271 Table 4: Direct and indirect costs of malaria outpatient care (GH¢) Cost Sub cost Number Sum total Mean Min Max category Category that GH¢(US$) GH¢(US$) GH¢(US$) GH¢(US$) paid OOP Direct cost Non-medical cost 323 3775.19(993.47) 11.69(3.08) 0.2(0.05) 108(28.42) Medical cost 140 1667.2(438.74) 11.91(3.13) 0.5(0.13) 140(36.84) Total direct Non-medical cost plus medical cost 366 5442.39(1432.21) 14.87(3.91) 0.2(0.05) 166(43.68) Indirect Indirect cost cost 207 5184(1364.21) 25.04(6.59) 6(0.16) 114(30) Overall Direct plus cost indirect cost 382 10626.39(2796.42) 27.82(7.32) 0.2(0.05) 200(52.63) 272
273
274
275 As presented in Table 5, the average cost incurred in the treatment of malaria by the uninsured was
276 greater (GH¢43.95/US$11.57) than their insured counterparts (GH¢24.75/US$6.51).
277 Table 5: Cost by insurance status
Insurance Number Sum Mean Min Max status that paid GH¢(US$) GH¢(US$) GH¢(US$) GH¢(US$) OOP Insured 321 7945.3(2090.87) 24.75(6.51) 0.2(0.05) 196.5(51.71) Uninsured 61 2681.09(705.55) 43.95(11.57) 0.5(0.13) 200(52.63) Total 382 10626.39(2796.42) 27.82(7.32) 0.2(0.05) 200(52.63)
278
279
14 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
280 Hospitalization
281 Out of the 393 patients interviewed, 13.2% (52/393) were admitted and incurred cost. The average
282 number of days spent by patients that were admitted was 4 days (range 2- 7days). The average
283 costs incurred by those who were admitted was GH¢18.94 /US$4.98(range: GH¢0.5/US$0.13 -
284 GH¢150/US$39.47). Surprisingly, the insured paid more hospitalization cost
285 (GH¢20.41/US$5.37) than the uninsured (GH¢14.55/US$3.83).
286
287 Number of days lost due to malaria
288 The average number of days a malaria episode prevented patients from working was about 5 days
289 for patients and 4 days for the caregivers. Controlling for the patients actively engaged in an
290 income generating activities, the number of days lost still remained at 5 days. The number of school
291 days lost by patients attending school due to malaria was 3 days (range: 1- 7 days).
292
293
294
295 Discussion
296
297 The main aim of the study was to examine the health seeking behaviours as well as cost of
298 treatment of malaria to households in the Kassena-Nankana district of Ghana in the face of multiple
299 first line antimalarial drugs.
300 This study revealed that a large proportion of malaria patients treated with ACTs (66.9%) and a
301 little more than 50% obtained the ACTs from formal health facilities (hospitals, Health centres,
302 clinics and CHPS). High utilization of formal health facilities for malaria treatment has been
15 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
303 reported in a previous studies[11–15]. The high use of formal health facilities in our study could
304 be attributed to NHIS as about 83% of patients were enrolled into the scheme. The positive effects
305 of health insurance enrolment on the utilization of formal health facilities has been widely
306 reported[12,16–19]. Although in our study, there was high utilization of formal health facilities for
307 ACT, buying of ACTs at the pharmacy/chemicals shops was also quite high (41.4%%) and
308 worrisome given that most of the pharmacy/licenced chemical shops do not carry out any
309 diagnostics test before treatment. In other studies conducted in Ghana, however, most people
310 reported visiting pharmacies and chemical shops to treat malaria than the formal health facilities
311 [20–22]. Some of the possible reasons for patient’s preference for chemical shops could be due to
312 waiting time at the health facilities and proximity[20].
313
314 Confirmation of malaria infection directs care to those in need. WHO recommends that all persons
315 from endemic areas with suspected malaria should be examined for evidence of infections with
316 malaria parasites by RDT or microscopy[1]. In this regard, the Roll Back Malaria (RBM)
317 partnership had set up a target that at least 80% of malaria patients are diagnosed and treated with
318 effective antimalarial medicines within 24 hours of the onset of illness[23]. Prompt diagnosis and
319 timely treatment can ease illness progression to severe stages and consequently decrease
320 mortality[24]. Given that in our study less than half of patients (49%) had diagnostic tests (Rapid
321 Diagnostic Tests or microscopy) to confirm malaria and only 23.6% took antimalarial within 24
322 hours of onset of illness is not encouraging and suggests that more efforts are still needed to
323 achieve the RBM target. Jima et al also reported a significant low percentage of respondents who
324 sought prompt care in Ethiopia[25]. People may delay before seeking care because of financial
325 constraint or because they want to be sure they are experiencing symptoms of malaria before they
16 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
326 seek care. A possible reason for the low diagnostic test before treatment could be due to limited
327 supply of RDTs at the health facilities[26] where majority of the study participants (53.6%) sought
328 care from.
329
330 Some of the reasons for the treatment failure of chloroquine included non-compliance with
331 duration of treatment regimen and misdiagnosis [27,28]. Therefore, diagnosis and rational use of
332 ACTs is crucial in the efforts to reduce malaria burden and to restrain the emergence of ACTs
333 resistance and treatment failure in the near future.
334
335 It is interesting to know that despite the fact that Artesunate-Amodiaquine (AA) and Artemether-
336 Lumefantrine (AL) are the most stocked ACTs in the market[29], this study revealed that the most
337 used ACT by patients in the Kassena Nankana District was Dihydroartemisinin-piperaquine
338 (36.1%) followed by Artesunate-Amodiaquine (35.4%) and Artemether-Lumefantrine (21.7%). In
339 fact, in terms of cost of antimalarials, Dihydroartemisinin-piperaquine is the most expensive ACT
340 sold in both the health facilities and the chemical shops in the district. For instance, whereas
341 Artesunate-Amodiaquine(AA) cost about GH¢5/US$1.3 in the pharmacies/chemical shops, P-
342 Alaxin (Dihydroartemisinin-piperaquine) cost about GH¢10/US$2.6, which is twice the price of
343 AA. Contrary to our findings, a previous qualitative study conducted in the same study area
344 reported that respondents preferred artemether-lumefantrine (AL) to other ACTs because AL was
345 perceived to be more effective and has minimal side effects as compared to the other ACTs
346 especially Artesunate-Amodiaquine[30,31]. In our study, a possible reason for the high reported
347 usage of Dihydroartemisinin-Piperaquine could be that patients were interested in trying a new
348 drug and also avoiding the side effects of AA. It also suggests that patients preferred
17 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
349 Dihydroartemisinin-Piperaquine because it has the best efficacy and profile in treating
350 uncomplicated malaria as observed in other studies [32]. Our study finding therefore supports the
351 importance of treatment options to take care of those who cannot bear AA and AL.
352
353 The overall average cost of treating malaria in the Kassena Nankana District was estimated at US
354 GH¢27.83/US$7.2. Though the cost of treatment is lower than malaria treatment cost estimated
355 in other similar studies conducted in other parts of Ghana [11,33], this amount is still relatively
356 high given that most of the respondents were from poor households. High treatment costs can
357 prevent the poor from seeking care when they have a fever episode as well as impoverish
358 households who seek care when they have an episode of fever [11,34].
359
360 The NHIS covers the treatment of about 95% of cases in Ghana including malaria. The benefit
361 package covers both inpatient and outpatient care, emergency and transfer services [35] and the
362 insured are generally not supposed to incur direct medical cost in treating malaria at any formal
363 health facility. In this study, the uninsured incurred more direct cost (GH¢14.91/US$3.9) than the
364 insured (GH¢10.87/US$2.9). Overall, the average cost incurred in the treatment of malaria by the
365 uninsured was greater (GH¢43.95/US$11.57) than their insured counterparts
366 (GH¢24.75/US$6.51). Thus, the uninsured paid US$5 more than the insured, which is consistent
367 with many findings where the uninsured paid more than the insured in treating malaria [12,15,36].
368 Though the insured were not supposed to incur direct medical costs, a possible reason for the cost
369 incurred by the insured could be due to co-payments for drugs, laboratory test, ward dues and other
370 informal payments. It could also be that antimalarial drugs were out of stock at the health facilities
371 and patient bought drugs outside the health facilities.
18 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
372
373 The average indirect cost (GH¢25.04/US$6.59) was 1.7 times higher than the direct cost
374 (GH¢14.87/US$3.91). This is consistent with many findings where the indirect costs of malaria
375 treatment were higher than the direct costs of treatment [11,15,33,37,38]. Therefore, efforts to
376 improve access to health care and reduction of financial burden to households in the treatment of
377 malaria should not only be directed to direct medical costs but also indirect costs.
378
379 Limitations
380 Recall bias is a possible limitation as respondents were asked to recall expenditure over a two
381 weeks period. The reported expenditure could either be overestimated or underestimated especially
382 in cases that respondents did not show receipts on expenditure but reported expenditure verbally.
383 However, because the recall period was short, we do not expect much recall issues to adversely
384 affect the study results.
385
386 In the study, we used self-reported fever or malaria to indicate malaria and we acknowledge that
387 as a limitation. For instance, not all fevers are malaria and also it is possible that some reported
388 malaria cases were really not malaria cases. Nevertheless, given that malaria is endemic in the
389 study area and mostly fever is associated with malaria, we assumed high knowledge about malaria
390 signs and symptoms by respondents would not affect the study findings much.
391
392
393
394
19 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
395 Conclusions
396 Though access to antimalarial medicine was quite good, prompt treatment such as treatment within
397 24 hours of onset of malaria as well as diagnostic testing before treatment of suspected fevers is
398 not encouraging when compared to the 80% target set by the Roll Back Malaria programme. There
399 was a positive effect of national health scheme enrolment on cost of treatment as the insured
400 incurred less cost (US$5 less) in treatment than the uninsured. Health facilities need to be
401 adequately equipped with malaria diagnostic equipment such as Rapid Diagnostic Tests (RDTs)
402 or microscopes and be encouraged to use them so as to direct limited and subsidized antimalarial
403 medicines to patients with malaria parasite infections. Also, more efforts are needed to improve
404 NHIS enrolment in order to reduce cost of treatment and accelerate universal health coverage.
405
406 List of abbreviations
407 AA Artesunate–Amodiaquine
408 ACTs Artemisinin-based Combination Therapies
409 AL Artemether-Lumefantrine
410 CHPS Community Based Health Planning and Services
411 DP Dihydroartemisinin-Piperaquine
412 DVD Digital Versatile Disc
413 HDSS Health and Demographic Surveillance System
414 INESS INDEPTH-Network Effectiveness and Safety Studies
415 IPTp Intermittent Preventive Treatment for pregnant women
416 IRS Indoor Residual Spraying
20 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
417 KND Kassena-Nankana District
418 LLINs long-lasting insecticidal nets
419 NHIS National Health Insurance Scheme
420 NHRC Navrongo Health Research Centre
421 OOP Out-Of-Pocket payment
422 PCA Principle Component Analysis
423 RBM Roll Back Malaria
424 RDT Rapid Diagnostic Test
425 WHO World Health Organization
426
427 Competing interests
428 The authors declare that there are no competing interests.
429 Authors’ contribution
430 MAD PA PDA PAD SC TA and AO: contributed in the conception, design, data collection,
431 analysis, interpretation, and drafting of the manuscript. All authors read and approved the final
432 manuscript.
433
434 Consent for publication
435 Not applicable
436 Availability of data and materials
437 Relevant data based on which conclusions were made are included in the document. The
438 questionnaire is included as supplementary material.
21 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
439 Funding
440 This research was funded by the Bill and Melinda Gate foundation and facilitated by the INDEPTH
441 Network
442
443 Acknowledgement
444 We wish to express our gratitude to the Navrongo Health Research Centre for the institutional
445 support to undertake the study. We are grateful to the study respondents, data collectors, data
446 managers and the data entry clerks for the various support they provided to this study.
447
448
449
450
451
452
453
454
455
456
457
458
459
460
22 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
461 References
462 1. WHO. World Malaria report [Internet]. 2014. Available: 463 http://www.who.int/mediacentre/factsheets/fs094/en/
464 2. Xu J-W, Xu Q-Z, Liu H, Zeng Y-R. Malaria treatment-seeking behaviour and related 465 factors of Wa ethnic minority in Myanmar: a cross-sectional study. Malar J. 2012;11: 417. 466 doi:10.1186/1475-2875-11-417
467 3. Ghana Health Service. National Malaria Control Programme [Internet]. 2014 [cited 29 Apr 468 2015]. Available: http://www.ghanahealthservice.org/ghs-subcategory.php?cid=4&scid=41
469 4. Maheu-Giroux M, Castro MC. Cost-effectiveness of larviciding for urban malaria control in 470 Tanzania. Malar J. 2014;13: 477. doi:10.1186/1475-2875-13-477
471 5. Yeka A, Harris JC. Treating uncomplicated malaria in children: comparing artemisinin 472 based combination therapies. Curr Opin Pediatr. 2010;22: 798–803. 473 doi:10.1097/MOP.0b013e32833fac44
474 6. World Health Organization, editor. Guidelines for the treatment of malaria. 2nd ed. Geneva: 475 World Health Organization; 2010.
476 7. Ghana Health Service. ANTI-MALARIA DRUG POLICY FOR GHANA [Internet]. 2009. 477 Available: http://apps.who.int/medicinedocs/documents/s18072en/s18072en.pdf
478 8. Oduro AR, Wak G, Azongo D, Debpuur C, Wontuo P, Kondayire F, et al. Profile of the 479 Navrongo Health and Demographic Surveillance System. Int J Epidemiol. 2012;41: 968– 480 976. doi:10.1093/ije/dys111
481 9. Oduro AR, Koram KA, Rogers W, Atuguba F, Ansah P, Anyorigiya T, et al. Severe 482 falciparum malaria in young children of the Kassena-Nankana district of northern Ghana. 483 Malar J. 2007;6: 96. doi:10.1186/1475-2875-6-96
484 10. Khatib RA, Selemani M, Mrisho GA, Masanja IM, Amuri M, Njozi MH, et al. Access to 485 artemisinin-based anti-malarial treatment and its related factors in rural Tanzania. Malar J. 486 2013;12: 155. doi:10.1186/1475-2875-12-155
487 11. Tawiah T, Asante KP, Dwommoh RA, Kwarteng A, Gyaase S, Mahama E, et al. Economic 488 costs of fever to households in the middle belt of Ghana. Malar J. 2016;15. 489 doi:10.1186/s12936-016-1116-x
490 12. Dalaba MA, Akweongo P, Aborigo R, Awine T, Azongo DK, Asaana P, et al. Does the 491 national health insurance scheme in Ghana reduce household cost of treating malaria in the 492 Kassena-Nankana districts? Glob Health Action. 2014;7. doi:10.3402/gha.v7.23848
493 13. Ameh S, Welaga P, Kabiru CW, Ndifon W, Ikpeme B, Nsan E, et al. Factors associated 494 with appropriate home management of uncomplicated malaria in children in Kassena- 495 Nankana district of Ghana and implications for community case management of childhood
23 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
496 illness: a cross-sectional study. BMC Public Health. 2015;15. doi:10.1186/s12889-015- 497 1777-3
498 14. Abokyi LN, Asante KP, Mahama E, Gyaase S, Sulemana A, Kwarteng A, et al. Use of 499 Antimalarial in the Management of Fever during a Community Survey in the Kintampo 500 Districts of Ghana. Carvalho LH, editor. PLOS ONE. 2015;10: e0142106. 501 doi:10.1371/journal.pone.0142106
502 15. Dalaba MA, Welaga P, Oduro A, Danchaka LL, Matsubara C. Cost of malaria treatment 503 and health seeking behaviour of children under-five years in the Upper West Region of 504 Ghana. PLOS ONE. 2018;13: e0195533. doi:10.1371/journal.pone.0195533
505 16. Fenny AP, Hansen KS, Enemark U, Asante FA. Quality of uncomplicated malaria case 506 management in Ghana among insured and uninsured patients. Int J Equity Health. 2014;13: 507 1–12. doi:10.1186/s12939-014-0063-9
508 17. Dalaba M, Akweongo P, Aborigo R, Ataguba J. To insure or not to insure:The influence of 509 insurance status on health seeking behaviour in the Kassena-Nankana district of Ghana. Afr 510 J Health Sci. 2012;21. doi:10.1093/heapol/czh031
511 18. Blanchet NJ, Fink G, Osei-Akoto I. The Effect of Ghana’s National Health Insurance 512 Scheme on Health Care Utilisation. Ghana Med J. 2012;46: 76.
513 19. Chankova S, Sulzbach S, Diop F. Impact of mutual health organizations: evidence from 514 West Africa. Health Policy Plan. 2008;23: 264–276. doi:10.1093/heapol/czn011
515 20. Ansah EK, Gyapong M, Narh-Bana S, Bart-Plange C, Whitty CJM. Factors influencing 516 choice of care-seeking for acute fever comparing private chemical shops with health centres 517 and hospitals in Ghana: a study using case–control methodology. Malar J. 2016;15. 518 doi:10.1186/s12936-016-1351-1
519 21. Awuah RB, Asante PY, Sakyi L, Biney AAE, Kushitor MK, Agyei F, et al. Factors 520 associated with treatment-seeking for malaria in urban poor communities in Accra, Ghana. 521 Malar J. 2018;17: 168. doi:10.1186/s12936-018-2311-8
522 22. Asante KP, Abokyi L, Zandoh C, Owusu R, Awini E, Sulemana A, et al. Community 523 perceptions of malaria and malaria treatment behaviour in a rural district of Ghana: 524 implications for artemisinin combination therapy. BMC Public Health. 2010;10. 525 doi:10.1186/1471-2458-10-409
526 23. Global Partnership to Roll Back Malaria. A decade of partnerships and results. Geneva: 527 World Health Organization; 2011.
528 24. Getahun A, Deribe K, Deribew A. Determinants of delay in malaria treatment-seeking 529 behaviour for under-five children in south-west Ethiopia: a case control study. Malar J. 530 2010;9: 320. doi:10.1186/1475-2875-9-320
24 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
531 25. Jima D, Getachew A, Bilak H, Steketee RW, Emerson PM, Graves PM, et al. Malaria 532 indicator survey 2007, Ethiopia: coverage and use of major malaria prevention and control 533 interventions. Malar J. 2010;9: 58. doi:10.1186/1475-2875-9-58
534 26. Boadu NY, Amuasi J, Ansong D, Einsiedel E, Menon D, Yanow SK. Challenges with 535 implementing malaria rapid diagnostic tests at primary care facilities in a Ghanaian district: 536 a qualitative study. Malar J. 2016;15. doi:10.1186/s12936-016-1174-0
537 27. Amponsah AO, Vosper H, Marfo AFA. Patient Related Factors Affecting Adherence to 538 Antimalarial Medication in an Urban Estate in Ghana. Malar Res Treat. 2015;2015: 539 e452539. doi:10.1155/2015/452539
540 28. Bloland PB, Organization WH, others. Drug resistance in malaria [Internet]. World Health 541 Organization Geneva; 2001. Available: 542 http://www.who.int/entity/csr/resources/publications/drugresist/malaria.pdf?ua=1
543 29. Tivura M, Asante I, van Wyk A, Gyaase S, Malik N, Mahama E, et al. Quality of 544 Artemisinin-based Combination Therapy for malaria found in Ghanaian markets and public 545 health implications of their use. BMC Pharmacol Toxicol. 2016;17. doi:10.1186/s40360- 546 016-0089-2
547 30. Chatio S, Aborigo R, Adongo PB, Anyorigiya T, Akweongo P, Oduro A. Adherence and 548 Uptake of Artemisinin-Based Combination Treatments for Uncomplicated Malaria: A 549 Qualitative Study in Northern Ghana. PLOS ONE. 2015;10: e0116856. 550 doi:10.1371/journal.pone.0116856
551 31. Chatio S, Aborigo R, Adongo PB, Anyorigiya T, Dalinjong PA, Akweongo P, et al. Factors 552 influencing adverse events reporting within the health care system: the case of artemisinin- 553 based combination treatments in northern Ghana. Malar J. 2016;15. doi:10.1186/s12936- 554 016-1172-2
555 32. Pekyi D, Ampromfi A, Tinto H, Coulibaly M, Tahita M, Victor M. Four artemisinin-based 556 treatments in African pregnant women with malaria. 2016;Malawi Medical Journal 28 (3): 557 11.
558 33. Zakaria H, Asante F. Economic Cost of Malaria Treatment under the Health Insurance 559 Scheme in the Savelugu-Nanton District of Ghana. Ghana J Dev Stud. 2015;10: 1. 560 doi:10.4314/gjds.v10i1-2.1
561 34. Ezenduka CC, Falleiros DR, Godman BB. Evaluating the Treatment Costs for 562 Uncomplicated Malaria at a Public Healthcare Facility in Nigeria and the Implications. 563 PharmacoEconomics - Open. 2017;1: 185–194. doi:10.1007/s41669-017-0021-8
564 35. Fenny AP, Hansen KS, Enemark U, Asante FA. Quality of uncomplicated malaria case 565 management in Ghana among insured and uninsured patients. Int J Equity Health. 2014;13. 566 doi:10.1186/s12939-014-0063-9
25 bioRxiv preprint doi: https://doi.org/10.1101/486183; this version posted December 3, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
567 36. Okoroh J, Essoun S, Seddoh A, Harris H, Weissman JS, Dsane-Selby L, et al. Evaluating 568 the impact of the national health insurance scheme of Ghana on out of pocket expenditures: 569 a systematic review. BMC Health Serv Res. 2018;18: 426. doi:10.1186/s12913-018-3249-9
570 37. Akazili J, Aikins M, Binka FN. Malaria treatment in Northern Ghana: What is the treatment 571 cost per case to households? Afr J Health Sci. 2008;14.
572 38. Hennessee I, Chinkhumba J, Briggs-Hagen M, Bauleni A, Shah MP, Chalira A, et al. 573 Household costs among patients hospitalized with malaria: evidence from a national survey 574 in Malawi, 2012. Malar J. 2017;16. doi:10.1186/s12936-017-2038-y
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
26