Transcript Poster presentation

Impact on prescribing patterns of a fee per
drug unit versus a fee per drug item
Kathleen Holloway1, Karkee SB2, Tamang AL2,
Gurung YB2, Pradhan R2, Reeves BC3
ICIUM 2004
1. Dept. Essential Drugs and Medicines Policy, WHO Geneva
2. Britain Nepal Medical Trust (BNMT), Kathmandu, Nepal
3. London School of Hygiene and Tropical Medicine, London
Background
• User fees are commonly charged to improve drug availability or modify
patient demand, but their effect on prescribing is rarely evaluated
• In Nepal, BNMT (an international NGO) operates community revolving
drug funds to supplement essential drugs in government health facilities
• Drug costs are shared between government, BNMT and patients; user fees
(about 40% of drug costs) are used to purchase more drugs
• Fee per prescription, 1-band fee per drug item and 2-band fee per drug
item were charged in 1995 in different districts, each fee type covering a
full course of treatment; previous research in this setting showed the
benefits of a fee per drug item compared to a fee per prescription, and no
difference between 1-band and 2-band fees per drug item
• All fee types were changed during 1996-2001 to a fee per unit (tab, cap) to
conform to government policy
Objectives
• To compare the effects of a fee per drug unit vs. a
fee per drug item (at similar below cost-price
levels) on prescribing quality and cost in
government primary health care facilities:
– Is charging per drug unit (40% of drug costs)
associated with the prescription of more
incomplete courses of drugs, compared with a
fee per drug item covering full courses of
treatment?
Study design
Non-randomised pre-post studies with controls
Comparisons in (1) 1995-2000 and (2) 2000-2
District
Taplejung
Bhojpur
1995
Fee per drug item
Fee per drug item
2000
Fee per drug item
Fee per unit (tab, cap)
2002
Fee per unit (tab, cap) Fee per unit (tab, cap)
Study design
Non-randomised pre-post studies with controls
Comparisons in (1) 1995-2000 and (2) 2000-2
District
Taplejung
Bhojpur
1995
Fee per drug item
Fee per drug item
2000
Fee per drug item
Fee per unit (tab, cap)
2002
Fee per unit (tab, cap) Fee per unit (tab, cap)
Study design
Non-randomised pre-post studies with controls
Comparisons in (1) 1995-2000 and (2) 2000-2
District
Taplejung
Bhojpur
1995
Fee per drug item
Fee per drug item
2000
Fee per drug item
Fee per unit (tab, cap)
2002
Fee per unit (tab, cap) Fee per unit (tab, cap)
Fee details
Fee type
Fee details (£1=NRs.80-110 during 1995-2002)
Fee per drug
item – 1 band
NRs.3 per item (HP) and NRs.5 per item
(hospital), whether cheap or expensive covering
a full course of treatment
NRs.5 per expensive item (antibiotics and
injections) and NRs.2 per cheap item (HP and
hospital) covering a full course of treatment
40% of the cost price of the drugs, charged by
the unit (tab, cap, bottle, inj). Drug prices were
listed in quantities equivalent to a full course
Fee per drug
item – 2 band
Fee per unit
Data Collection
• Prescribing indicators 1995, 2000 and 2002
– carbon copy prescriptions
• Drug availability 1995-2002
– Routine monitoring at health facilities
– >90% of key drugs available in all districts in all years
Sample sizes
• 10-12 facilities per fee type (one fee type / district)
• average 200 prescriptions per facility per year
– systematic random selection
– >30 prescriptions per facility per year for all indicators
1995-2000: fee per drug unit vs. fee per item
Px indicator
Av.no.item/Px
Fee per item (1995) vs.
Fee per unit (2000)
1995
2000
Diff.
2.1
2.2
+0.1
Fee per drug item
(1995 and 2000 - control)
1995
2000
Diff.
1.8
1.9
+0.0
% Px with AB
54.4
55.2
+0.8
54.6
53.4
–1.2
% Px with Inj
17.6
5.4
–12.1
15.3
10.9
–4.4
% Px with vits,
tonics
15.8
14.8
–1.0
8.4
10.7
+2.3
Av.no.units per
drug item
14.4
12.6
–1.8
15.7
14.4
–1.3
% AB pres. in
underdose
15.8
26.6
+10.8
11.4
14.8
+3.4
Av.cost/Px
23.7
30.5
+6.8
26.0
30.8
+4.8
2002-2000: fee per drug unit vs. fee per item
Px indicator
Av.no.item/Px
Fee per item (2000) vs.
Fee per drug unit (2002)
2000
2002
Diff.
1.9
1.8
–0.1
Fee per drug unit
(2000 and 2002 - control)
2000
2002
Diff.
2.2
2.4
+0.2
% Px with AB
53.4
52.4
–1.0
55.2
62.6
+7.4
% Px with Inj
10.9
7.2
–3.8
5.4
5.3
–0.1
% Px with vits,
tonics
10.7
9.0
–1.7
14.8
22.0
+7.2
Av.no.units per
drug item
14.4
13.7
–0.7
12.6
12.8
+0.2
% AB pres. in
underdose
14.8
19.8
+5.0
26.6
26.9
+0.3
Av.cost/Px
30.8
28.2
–2.6
30.5
33.5
+3.0
Estimates of effect of changing from fee per
drug item to fee per drug unit
Px indicator
Comparison 1: 1995/2000 *
Estimate
95% CI
Comparison 2: 2000/2002 †
p
Estimate
95% CI
p
Av.no.item/Px
0.03
-0.25 to 0.30
0.84
-0.29
-0.64 to 0.06
0.10
% Px with AB
1.86
-2.97 to 6.69
0.43
0.63
-6.00 to 7.26
0.84
% Px with Inj
-6.03
-4.10
-7.52 to -0.68
0.02
% Px with vits,
tonics
-4.53
-12.04 to 2.98
0.22
-3.18
-13.95 to 7.59
0.54
Av.no.units per
drug item
-1.71
-3.12 to -0.30
0.02
-1.68
-3.17 to -0.19
0.03
% AB pres. in
underdose
10.86
1.58 to 20.13
0.02
11.97
2.02 to 21.92
0.02
1.96
-0.95 to 13.87
0.73
-4.07
-13.42 to 5.27
0.37
Av.cost/Px
-9.59 to -2.46 0.002
* PxIndicator2000 = a + b1(Pxindicator1995) +b2(fee per unit) + error
† PxIndicator2000 = a + b1(Pxindicator2002) –b2(fee per item) + error
Conclusions
• Fees per unit as compared with fees per item covering a full
course of treatment were associated with:
–
–
–
–
the prescription of fewer injections
the prescription of slightly lower number of units per drug prescribed
the prescription of more antibiotics in under-dose
similar prescribing patterns with regard to the number of drugs per
patient, the drug costs per patient and the % of patients receiving
antibiotics and vitamins
• An EDL change and decentralisation of purchasing authority to
health facilities during the study may account for decreased use
of essential drugs and increased cost per prescription, so
masking prescribing and cost changes due to the fee systems.
Key lessons, policy implications and
future research
Key lessons
• Fee per unit, as compared to fee per drug item (covering a full course),
is associated with the prescription of fewer units per drug item and less
use of expensive items such as injections
Policy implications
• Revolving drug funds and insurance systems often charge per unit and
should beware adverse effects on prescribing quality particularly if the
fee level is unaffordable to patients
Future research
• Impact on prescribing quality of different fee types, set at different
levels, to ascertain (1) whether similar effects to this study are seen, and
(2) at what level of fee the positive effects of a fee per unit and a fee per
drug item are maximised (and the negative effects minimised)