Study participants were recruited as illustrated in the Flow Diagram (see figure 2), yielding (per design) 400 medicine shops treating infants <2months of age, 68 dispensing for such cases but not involved in assessment and treatment decision-making, and 33 not providing any services for young infants but potentially providing services for sick older infants and children. The sample also included 82 clinics involved in treatment of sick young infants, and four not involved in young infant care.
Service Provider Characteristics
Most medicine shops in urban settings, close to a hospital, were registered with the DDA; most of those more than 30 minutes away from a hospital were not (p<.001) (see table 3). Over three quarters of shops were staffed by individuals reporting at least two academic years of professional paramedical training (69% reported having credentials as certified medical assistants or health assistants). Pharmacists/ pharmacy assistants were more common in urban medical shops, close to a hospital (10%, vs. 6% in more distant shops, p=0.13). Most medicine shop providers were male (86%), as were almost all clinic-based physicians treating sick young infants (98%). Mean age of medicine shop providers and clinic-based physicians was the same (38 years). Forty-six percent of medicine shop providers reported more than 10 years of experience treating sick young infants (vs 31% of physicians, p=0.019). Fifteen percent of medicine-shop providers also worked in public sector health facilities; among physicians, the proportion was more than twice as high (34%; p=.002 on the difference). Almost all medicine shops reported having services available seven days a week (95%, vs 83% of clinic-based physicians, p=0.002) and being open at least 11 hours a day (93%, vs. 85% of physicians, p=0.1). A small proportion of shops (10%) reported that physicians were sometimes available at the shop to see patients; for those that did, typically this was once or twice a week (data available on request).
Table 2: Profile of medicine shops & private clinics, engaged in treatment of young infants
|
Medicine shops (%)
|
Physician-run clinics (%)
|
Proximity to hospital with in-patient pediatrics service (in minutes)
|
<30 min
n=200
|
30–60 min
n=100
|
>60 min
n=100
|
All
n=400
|
n=82
|
DDA registration
|
74
|
36
|
34
|
55
|
N/A
|
Professional credentials:
|
|
|
|
|
|
Pediatrician
|
|
|
|
|
38
|
Other physician
|
|
|
|
|
62
|
CMA/ HA
|
66
|
74
|
72
|
69
|
|
Pharmacist/ Pharmacy Assistants
|
10
|
6
|
5
|
8
|
|
ANM/ nurse
|
6
|
5
|
7
|
6
|
|
Other paramedical
|
1
|
2
|
6
|
3
|
|
No professional training
|
17
|
12
|
10
|
14
|
|
Sex (% male)
|
86
|
94
|
79
|
86
|
98*
|
Age - mean
|
|
|
|
38 years
|
38 years
|
<30 years
|
22
|
20
|
30
|
24
|
17
|
30 to 40 years
|
44
|
52
|
43
|
45
|
57
|
>40 years
|
34
|
28
|
27
|
31
|
26
|
10+ yrs experience treating sick infants
|
53
|
34
|
43
|
46
|
31†
|
Dual practice in public sector H facility
|
13
|
17
|
19
|
15
|
34*
|
Also does in-patient pediatrics
|
|
|
|
|
62
|
Services available:
|
|
|
|
|
|
11+ hours/ day
|
94
|
88
|
97
|
93
|
85
|
7 days/ week
|
98
|
96
|
90
|
95
|
83*
|
Physician on site at least once/ week
|
9
|
5
|
11
|
9
|
N/A
|
Notes: Medicine shop data presented in this table are restricted to those to which the full survey instrument (Tool 1) was administered (n=400); shops that reported only dispensing medicines for young infant illness and not involved in assessment and treatment decisions are not included here.
Acronyms: DDA = Department of Drug Administration; CMA = Certified Medical Assistant; HA = Health Assistant; ANM = Auxiliary Nurse-Midwife; H facility = health facility; N/A = not applicable.
* p-value on difference between medicine shops and clinics <0.01
† p-value on difference between medicine shops and clinics <0.05
Among those only dispensing (n=68)—which are not included in table 2—the profile is similar, with the following exceptions: 88% were in locations within 30 minutes of a hospital providing in-patient care (vs. 50% among those involved in assessment and treatment, p-value on the difference = 0.001), 93% were DDA-registered (vs. 55%, p-value on the difference <0.001), and 24% were staffed by pharmacists or pharmacy assistants (vs. 8%, p<0.001 on the difference).
Infant and Child Illness-Related Services Offered
Among medicine shop-based practitioners who reported treating infants <2 months of age, most (56%) reported having treated no more than 20 cases with oral antibiotics over the previous six months. A relatively small proportion (10%) reported much higher volumes, over 100 cases. The volume of such cases reported by physicians was, on average, considerably higher (data available on request).
Although only a comparatively small proportion of medicine shop practitioners (19%) reported having used parenteral antibiotics for treating young infants over the past 6 months, the proportion was considerably higher among those based more than an hour away from a hospital (33%) compared to those within 30 minutes (13%, p=.027 on the difference). Almost half of the physicians in the sample reported having used parenteral antibiotics for at least some of their young infant patients (46%), although the proportion reporting such treatment for infants <1 month of age was smaller (36%).
By design, as indicated earlier, the survey was to include 400 medicine shop-based practitioners who reported treating infants <2 months of age. This practice was determined through use of a screening instrument. In administering the instrument:
- an additional 68 were identified, who reported only dispensing for infants of this age (not assessing and treating)—they were administered a somewhat shorter survey instrument (Tool 2); and
- a further 33, who reported no treatment or dispensing for young infants.
All of those administered the screening instrument (n=501) were asked a small set of questions concerning services for illness among older infants and children, aged two to 59 months. As seen in table 3, a large proportion reported that they were involved in assessing and treating diarrhea (87%) and acute respiratory infections (86%) in this age group, although the proportion reporting such a service was smaller among shops close to hospitals (79%) than those more distant (for diarrhea, p-value on the difference <0.001; for ARI, p-value=0.007).
Table 3: Medicine Shops Treating Diarrhea & ARI, ages 2-59 months, vs. only Dispensing
Proximity to hospital (minutes)
|
<30 min (%)
n=279
|
30-60 min (%)
n=114
|
>60 min (%)
n=108
|
All (%)
n=501
|
Diarrhea
|
79
|
98
|
95
|
87
|
Acute respiratory infection
|
79
|
95
|
94
|
86
|
Readiness for Treatment of Sick Young Infants (n=400 medicine shops, n=82 clinics)
Only 15% of medicine shops providing treatment for sick young infants had a scale suitable for weighing young infants vs. 72% of medical clinics (p<0.001); 74% had an adult scale vs. 95% of clinics (p<0.001). (By comparison, in a nationally representative survey of public sector health facilities—the 2015 Nepal Health Facility Survey (NHFS)—infant scales were found in 65% of health posts [19]). Digital thermometers were present in 94% of medicine shops and 98% of clinics (and 94% of health posts in the NHFS). Similarly, stethoscopes were present in 97% of medicine shops and 100% of clinics (and 100% of health posts, NHFS), and 94% of medicine shop-based practitioners had a cell phone (vs. 100% of physicians). Eighty-five percent of medicine shops had a timer or watch available for counting respiratory rate, vs. 99% of clinics (95% of health posts in NHFS). Pulse oximeters were present in 16% of medicine shops and 76% of clinics.
Commodities found to be available at the time of the survey in almost all medicine shops included ORS sachets (98%, although only in 88% among shops in mountain districts, p=0.013), pediatric formulations of amoxicillin (97%), and cefixime (93%). Zinc was available in over three quarters of medicine shops in hill and plains districts but only 43% of shops in mountain districts (74% in the sample, as a whole; p-value on the difference <0.001). Cotrimoxazole suspension or dispersible tablets were somewhat less widely available: 62% of shops in plains, 57% hill, and 49% of mountain districts. Injectable antibiotics suitable for treating pneumonia or sepsis were available in fewer of the shops (ceftriaxone 55%, gentamicin 40%, cefotaxime 37%, and ampicillin 24%).
Just over a quarter of medicine shop-based practitioners (27%) reported ever having received IMNCI training, with a higher proportion among those within 30 minutes of a hospital (33%, p-value on the difference = 0.24). Half of surveyed physicians reported having received this training (49%, p=0.013 on the difference with medicine-shop providers).
Reference materials present on the treatment of sick infants and children included: Current Index of Medical Specialties (42% of medicine shops, 40% of clinics), treatment guidelines for the government’s IMNCI program (25% of medicine shops, 50% of clinics (and 67% of health posts in NHFS [19])), and course books (8% of medicine shops, 24% of clinics). Registers were present and used for recording information on cases of sick infants in 12% of medicine shops and 57% of clinics.
Quality of Care for Sick Young Infants
A variety of open-ended questions (without specific prompts) were asked related to quality or appropriateness of care for sick young infants. For several potential danger signs, fewer medicine shop practitioners than physicians reported specifically considering them when assessing sick young infants, notably: level of consciousness, umbilical redness or pus, and seizures or feeding problems as reported by the care-giver (see table 4). The survey also included an unprompted question on what assessment findings would suggest that a sick young infant may have a potentially severe infection. Most medicine shop providers responded that high temperature (89%), severe chest in-drawing (82%), and rapid respiratory rate (86%) would be danger signs. Other important potential danger signs that were less frequently mentioned included: abnormally low temperature (38%), care-giver report of poor feeding (29%) or seizures (20%), and moving only when stimulated (17%) or unconscious (14%). A larger proportion of clinic-based physicians were able to cite at least four recognized dangers signs of possible severe infection (81%), than medicine shop-based practitioners (66%) although because the questions were asked slightly differently, we have not calculated a p-value on this apparent difference.
Most medicine shop providers and physicians reported using amoxicillin as first-line oral antibiotic for these cases (in line with national guidelines), followed by cefixime (see table 4, below). Choice of first-line injectable antibiotics appeared to differ, with physicians reporting ampicillin (37%) and cefotazime (29%) or ceftriaxone (24%); and medicine shop providers reporting gentamicin (53%), although in both cases the samples were small.
The same proportion of medicine shops and clinics reported commonly using bronchodilators for treating sick young infants (43%). In most cases, this would be inappropriate. Eleven percent of medicine-shop providers reported at least some use of injectable steroids for treating sick young infants; 6% within the past 6 months (with no differences in proportion across proximity strata); 21% of physicians reported at least some use of steroids, 12% over the previous six months (for the difference between medicine shops and clinics in any reported use of steroids, p-value = 0.013).
Over one third of medicine shop providers (35%) reported determining antibiotic dosing for young infants based on age, not weight; this was less common among physicians (10%, p-value on difference <0.001). Of those who reported determining dosage based on weight, only 10% of medicine shops reported using a suitable scale (Salter or pan) vs. 63%, among physicians (p-value <0.001). Over three quarters of medicine shop providers (79%) used a technique that entailed weighing the mother with and without the baby and subtracting to determine the baby’s weight. This was also commonly done by physicians (37%, p-value < 0.001). Almost all medicine shop practitioners (97%) reported that they do not normally remove the baby’s clothing/ coverings to do the weighing; similarly, 80% of physicians also reported not removing the baby’s clothing/ coverings (p-value on the difference = 0.005). Such practices compromise accuracy of antibiotic dosing (and safety, notably for aminoglycoside antibiotics).
Giving an abbreviated course of antibiotics was done at least somewhat often by close to half of medicine shop providers (48%) and physicians (40%). A significant proportion of both medicine shop providers and physicians reported commonly helping arrange transport for referred cases (65% and 59%, respectively, p-value = 0.37) and providing a referral note (52% and 77%, respectively; p-value = 0.001). Only 22% of medicine shop providers reported calling the physician at the receiving institution (vs. 39% of the physicians, p-value on the difference = 0.02).
Table 4: Quality/ Appropriateness of Care for Sick Young Infants (unprompted questions)
|
Medicine shops (%)
|
Physician-run Clinics (%)
|
Proximity to hospital (in minutes)
|
<30 min
|
30–60 min
|
>60 min
|
All
|
|
Sick young infants <2 months of age
|
n=200
|
n=100
|
n=100
|
n=400
|
n=82
|
Reports assessing for:
|
|
|
|
|
|
Respiratory rate
|
92
|
92
|
86
|
91
|
90
|
Temperature
|
91
|
80
|
90
|
88
|
89
|
Feeding (as reported by mother)
|
67
|
65
|
61
|
65
|
82†
|
Seizures (as reported by mother)
|
17
|
36
|
23
|
24
|
62†
|
Weight
|
28
|
44
|
29
|
32
|
66†
|
Chest in-drawing
|
48
|
59
|
57
|
53
|
60
|
Umbilical redness or pus
|
31
|
25
|
24
|
28
|
48††
|
Level of consciousness
|
14
|
21
|
23
|
18
|
46†
|
Treatment
|
|
|
|
|
|
Usual first-line oral antibiotic
|
n=200
|
n=100
|
n=100
|
n=400
|
n=82
|
Amoxicillin (+/- clavulanate)
|
63
|
76
|
77
|
73
|
82
|
Cefixime
|
41
|
30
|
28
|
35
|
39
|
Cefpodoxime
|
1
|
7
|
2
|
3
|
17
|
Cotrimoxazole
|
6
|
13
|
8
|
8
|
8
|
others
|
7
|
8
|
18
|
10
|
8
|
Usual first-line injectable antibiotic*
|
n=28
|
n=17
|
n=36
|
n=81
|
N =38
|
Gentamicin
|
51
|
34
|
63
|
53
|
21††
|
Ampicillin
|
16
|
9
|
16
|
14
|
37††
|
Cefotaxime
|
20
|
47
|
16
|
24
|
29
|
Ceftriaxone
|
20
|
17
|
22
|
20
|
24
|
Amikacin
|
9
|
20
|
0
|
7
|
16
|
others
|
0%
|
0
|
3
|
1
|
11††
|
Other treatments used
|
n=200
|
n=100
|
n=100
|
n=400
|
N=82
|
Bronchodilators
|
50
|
37
|
34
|
43
|
43
|
Injectable steroids
|
11
|
11
|
11
|
11
|
21
|
Steroids given within past 6 mo.
|
5
|
10
|
7
|
6
|
12
|
Dosage determination & weighing
|
n=200
|
n=100
|
n=100
|
n=400
|
N=82
|
Determines dose by age, not weight
|
30
|
30
|
50
|
35
|
10†
|
Doses by weight, determined by**:
|
n=140
|
n=68
|
n=50
|
n=258
|
N=71
|
Salter or pan scale
|
9
|
9
|
12
|
10
|
63
|
Adult scale (subtract. technique)
|
82
|
81
|
68
|
79
|
37
|
Estimates by looking
|
9
|
10
|
20
|
11
|
0%
|
Among those weighed,
|
n=129
|
n=61
|
n=39
|
n=229
|
N=71
|
leaves baby’s clothes on
|
97
|
97
|
100
|
97
|
80††
|
Shortened treatment course
|
n=200
|
n=100
|
n=100
|
n=400
|
N=82
|
Somewhat or very often
|
50
|
53
|
36
|
48
|
40
|
Danger sign referral
|
n=200
|
n=100
|
n=100
|
n=400
|
N=82
|
Helps arrange transport
|
60
|
83
|
57
|
65
|
59
|
Provides referral note
|
42
|
65
|
58
|
52
|
77†
|
Calls ahead to MD at receiving HF
|
25
|
16
|
20
|
22
|
39††
|
Gives pre-referral oral antibiotics
|
42
|
60
|
59
|
51
|
48
|
Gives pre-referral inj. antibiotics
|
5
|
8
|
13
|
8
|
12
|
Schedules follow-up visits
|
99
|
98
|
99
|
99
|
95
|
* denominator includes only those reporting having used injectable antibiotics to treat sick infants over the previous 6 months.
** denominator includes only those reporting determining dose based on weight
† p-value on difference between medicine shops and clinics <0.001
†† p-value on difference between medicine shops and clinics <0.05
Quality of Care for Older Infants and Children up to 59 months
For diarrhea, most medicine shops were doing more than dispensing medicines; 87% reported assessing and making treatment decisions (97% of those located ≥30 minutes from a hospital, vs. 79% among those <30 minutes, p<0.001). In table 5, we present reported practices restricted to those doing more than dispensing. Dispensing oral rehydration solution (ORS) and zinc—over-the-counter products—falls within their legally permitted scope of practice. Overall, the likelihood of appropriate treatment for diarrhea was lower in medicine shops than in physician-run clinics. The proportion who reported prescribing ORS most or all of the time was 91% (vs. 98% of clinics, p-value = 0.06); 66% of medicine shops reported routinely using zinc (vs. 90% of clinics, p-value on the difference < 0.001). A small proportion of providers reported routinely using antibiotics to treat diarrhea (6% of medicine shops, 4% of physician-run clinics), however 76% of medicine shop practitioners and 70% of physicians reported at least some use of antibiotics for non-bloody diarrhea (p-value on the difference = 0.13). Metronidazole was the most commonly used antibiotic for non-bloody diarrhea (medicine shops 75%, clinics 69%, p=0.22). Other than for specific clinical presentations, e.g. suggestive of giardiasis, this is an inappropriate antibiotic for diarrhea.
Table 5: Quality of Treatment for Diarrhea & ARI, among Infants/ Children 2-59m
|
Medicine shops (%)
|
Physician-run Clinics (%)
|
Proximity to hospital (in minutes)
|
<30 min
|
30–60 min
|
>60 min
|
All
|
Diarrhea
|
|
n=222
|
n=112
|
n=103
|
n=437
|
n=86
|
ORS most/ all cases
|
93
|
89
|
88
|
91
|
98
|
Zinc most/ all cases
|
69
|
61
|
67
|
66
|
90†
|
No antibiotics for non-bloody diarr
|
25
|
23
|
22
|
24
|
30
|
Antibiotics for bloody diarr:
|
|
|
|
|
|
Ciprofloxacin or other quinolone
|
19
|
40
|
32
|
32
|
18††
|
Metronidazole
|
26
|
28
|
29
|
27
|
31
|
A cephalosporin antibiotic
|
16
|
11
|
11
|
14
|
38†
|
Cotrimoxozole
|
22
|
18
|
23
|
21
|
11
|
Acute Respiratory Infection
|
|
n=220
|
n=109
|
n=102
|
n=431
|
n=86
|
Antibiotic based on respir. rate
|
97
|
99
|
98
|
98
|
98
|
Specific antibiotics used:
|
|
|
|
|
|
Amoxicillin +/- clavulanate
|
70
|
68
|
70
|
69
|
65
|
Cefixime
Other cephalosporin
Cotrimoxazole
Azithromicin
|
17
3
8
0
|
23
4
5
0
|
16
2
8
0
|
18
3
7
0
|
12
9
2
7
|
Other
|
3
|
1
|
5
|
3
|
5
|
† p-value on difference between medicine shops and clinics <0.001
†† p-value on difference between medicine shops and clinics <0.05
For bloody diarrhea, one third of medicine shop practitioners (32%) reported routinely giving ciprofloxacin or other fluoroquinolone antibiotics as first line treatment (in line with IMNCI guidelines) but only 18% of clinics (p-value on the difference = 0.016). Metronidazole was commonly used (reported by 27% of medicine shops, 31% of clinics, p-value 0.32); oral cephalosporins were reported by 38% of clinic-based practitioners as first line. Neither are considered appropriate treatment.
For ARI (as with diarrhea case management), the overwhelming majority of medicine shops (86%) reported not just dispensing treatment but also assessing and making treatment decisions. Among semi-proximal and remotely located medicine shops, 94% reported assessing, making treatment decisions, and dispensing. Since this entails dispensing of antibiotics without a physician’s prescription, this lies outside the formally recognized scope of practice for non-physicians working in the private sector.
Virtually all providers reported using respiratory rate to classify ARI cases for antibiotic treatment. Amoxicillin (+/- clavulanate) was reported as first-line treatment by most providers in medicine shops (69%) and physician-run clinics (65%), consistent with the government’s IMNCI recommended treatment. Cefixime was the second most often reported first-line antibiotic (18% of medicine shops, 12% of clinics). Although not the recommended first-line drug, this is also an efficacious treatment for ARI. Overwhelmingly, providers reported using syrup/ suspension formulations for treating young children (96%), not dispersible tablets (which typically are used in public-sector programs).