Cost and performance of malaria surveillance in Thailand

0277-Y536.‘84 $3 00 + 0.W Pergamon Press Ltd

COST AND PERFORMANCE OF MALARIA IN THAILAND SOMKID ‘Facult>

of Economics.

KAEWSONTHI’

Chulalongkorn Bradford,

SURVEILLANCE

and ALAN G. HARDING’

University, Bangkok, Thailand and West Yorkshire BD7 IDP. England

‘University

of Bradford.

AbstractSome results arc presented from a study to determine the costs and performance of the antiparasite elements of the malaria disease control programme in Thailand. Issues examined in the paper are the concept of cost-effectiveness and its relevance in evaluating health care processes. procedures for measuring the performance of malaria surveillance and monitoring processes and procedures for measurrng costs. Some data on the costs and the performance ofoperational services and control activities in two malaria zones are presented and the paper closes by considering how health economics research can be stimulated and supported in developing countries. It is argued that the cost-effectiveness of malaria surveillance and monitoring processes (and probably man!, other health care processes) can not be measured retrospectively. In the case of malaria control the effectiveness ofeach operational services can not be compared because services provided are interactive and complemcntar! rather than alternatives. In addition the targets set and levels of effectiveness achieved may not be the same for these complementary processes. Procedures developed for measuring the performance and costs of malaria operational services and activities are described. Five types of measurement are used to evaluate the performance of the operational services and activities: efiectlveness ( I;,,);time (days); performance ( :,,); and efficiency (cost/unit). Actual expenditure on malaria operational services and activities is not known since all expenditure of government departments in Thailand is recorded under nine budget headings. Budget expenditure at division. region and zone levels must therefore be apportioried to assess the costs of operational services and activities. Since a variety of criteria may legitimately be used to apportion costs at each level, a network technique was developed which allows examination of the effect of all possible combinations of criteria. By this means the maximum, minimum and most appropriate costs are determined. Examples of the costs and performance of surveillance operational services and activities in two zones are presented. Data illustrates the outcomes from the procedures developed and indicates how malaria disease control managers might interpret and use the information obtained. The paper closes with some observations on how health economics research in developing countriescould be better stimulated and supported through staff development programmes and through supervised ‘on the job’ training.

INTRODUCTION

In line with the objectives of the conference the paper presents some of the results of research to determine the costs and performance of the antiparasite elements of the malaria control programme in Thailand. Consideration is also given to how health economics research might be better stimulated and supported. Control of the prevalence of malaria cases in Thailand is achieved through the antiparasite and antivector measures of the malaria surveillance and monitoring programme (Fig. I ). Malaria surveillance embraces a number of activities; the detection of malaria infection in the community by screening all people reporting a fever; parasitological examination of their blood: antimalarial drug treatment; epidemiological investigation and. where appropriate, implementation of remedial and/or preventive measures such as mass dru+ administration and insecticide spraying (Fig. 2). Monitoring is the system developed for recording the outcome from surveillance activities (Fig. 3). The study was only concerned with the antiparasite activities of surveillance and monitoring; collection of blood slides from patients with a recent history of fever: microscopic examination of the blood slide to determine if malaria parasites are present in the blood and to identif) the type of parasite present; presumptive treatment to provide some relief from

symptoms and to render the patient non-infective to anophelines for approx. IO days and/or radical treatment; case investigation to determine where infection occurred and possible transmission; followup (in principle) of positive cases for 1 year to check that patients are clear of infection; and monitoring, recording the results from all stages of surveillance. Collection of blood slides and other antiparasite activities are undertaken by seven types of operational service. Activities undertaken by each operational service are shown in Table 1. Patients may visit centres for investigation; malaria clinics (MC); malaria village volunteers (MVV); village health centres (VHC); hospitals (H); and mobile clinics. This is termed passive case detection (PCD). In some areas PCD may be supplemented by a system of regular domiciliary visits termed active case detection (ACD). In 1980, $17.3 x IO6 was spent on the malaria control programme in Thailand. While malaria disease control managers are conscious of the need to obtain the best return from the money expended there is little guidance available on how they might evaluate and thereby seek to improve their activities. Aims initially set for the research were: (i) to determine HOW the cost-effectiveness of malaria surveillance and monitoring processes could be measured and the results used in decision-making by malaria disease control managers; and 1081

SOMKID KAEWSONTW

1082

--.J

and

ALAN

6.

HARDING

OffIce

of work

office

office

Regional

Division

Zone qffice

Sec1or

Locotton

visitor

report

I

Sector

service

I

(division)

t

cl),

chief

Fig. 2. Flowchart

of

of case detectlqn

Summary report surveillance (l),(2),(3),(4), (5),(6).(7),lEt)

Summery

so CH

7

House

I-Malaria

Case detectlon

(31,

(4),

(51

1

,

c

operation

1

I Anolysed 3 month report from region ond division

Head of EPI section

I

(8)

investigation

Focal

RegIonal director

Report of CI + RT

Report of CI + RT

Case lnvestigatlon and RT by sector chief assistant, sector chief or

(7)

Analysed 3 month report from zone and region

16)

(6)

servtce

: Epidemiologicol study Radical treo?ment. FI

Summary of surveillance. A 3 month report Anolysii of 3 n?or Ilh summary report 1 Report of case in! ‘ected in none spray area

(5),

r Medical

of surveillance

3 month performance

(21,

I

1

Case ,nves,iga+ion

Head

Region

Zone

sector

Field

quarter

monthly

MS10

(3)

II)

on

j

MS3(4)

-Iocotlo” mopping

performance

“i”3”‘MSl~l~~~mo”thy

MS2

hlSlbl4)

work

-LL

1 Reporting

MS1

(I

HQs

EPI

I

MS3

EPI

I

Sector

Reglon

monthly

w!th

section

I MS5

section

(4)

monthly

I

I

Summary Of surveillance

monlly

with

sltde

(I)

MS5

c hlrf

blood

MS4

eY

’

on

HOS

(4)

EPI

EPI

I

Sector

Region

MS6

Actlon

MS?

(II)

MS6o

sectIon

MS6

-l

reglstrotlon

sec11on

I

chief

and

cases

Focus

of

(4)

mvesttgotion

every

MS6o

foct

i”vesUgatlo”

and

(ii) to strengthen the research capability in health economics at Chulalongkorn University. Bangkok, Thailand. Since it was found that the cost-effectiveness of malaria control processes cannot be measured retrospectively the first aim was, of necessity, amended to HOW to determine the COSTS and the PERFORMANCE of surveillance and monitoring processes. Five areas are therefore examined in the paper; the concept of cost-effectiveness and its relevance in the evaluation of health care processes: procedures for measuring the performance of surveillance and monitoring processes; procedures for measuring costs; some data on the costs and performance ofoperational units in two malaria zones; and how health economics research could be better stimulated and supported in developing countries.

These conditions are not met in a retrospective study of surveillance and monitoring processes nor. probably. in the evaluation of many other health care processes. Operational services draw upon the same population and operate in the same area at the same time. The services are complementary and not alternatives. Operational services in different areas will have markedly different conditions, may not have the same target and are unlikely to achieve the same level of effectiveness. Surveillance activities undertaken by operational services such as blood slide collection and examination are not alternatives but essential elements. Cost-effectiveness of operational services could be measured in carefully planned and controlled experimental conditions. But it is doubtful if such costly experiments would yield results of real value in practical situations.

COST-EFFECTIVENESS

MEASUREMENT OF PERFORMANCE

The research was mounted with the intention of comparing the cost-effectiveness of surveillance processes (operational services and activities) within an area and comparing the cost-effectiveness of similar processes in different areas. (A process is defined as a particular combination of inputs or activities and effectiveness defined as the extent to which a target is achieved.) The concept of cost-effectiveness, determining the costs of processes at the same level of effectiveness in order to recommend the best course of action is intellectually attractive and deceptively simple. But the cost-effectiveness of surveillance and monitoring processes cannot be measured retrospectively. To determine which of two or more processes is the more cost effective at least three conditions must be satisfied. (i) Processes must be real alternative and not complementary. (ii) Processes must have the same target and achieve the same level of effectiveness. (iii) Environmental conditions should be similar.

Since cost-effectiveness could not be used to evaluate the performance of surveillance activities and operational services other forms of criteria were used; effectiveness of each process (:,). time (days); performance (7;); relative contribution ( y,); and efficiency expressed as cost/unit. Performance criteria for operational services and activities are shown in Tables 2 and 3 respectively. Eflectiveness

( Oo)

Percentage effectiveness measures the extent to which a target is met, e.g. if a target is to provide treatment within 5 days, 98 % effectiveness would indicate that 2 o,0 of cases took longer than 5 days.

Performance of some activities is best expressed as the time taken to complete the activity. Performance may be expressed as the average time (arithmetic mean and standard deviation) and the cumulative percentage completed within successive days.

SOMKID KAEWSONTHI and ALAN

1086 Performance

(“,)

Percentage performance expresses the degree to which a task is successfully completed, e.g. the percentage of malaria village volunteers who are active. Relatice

contriburion

(3,)

Percent relative contribution expresses the contribution made by each operational service to surveillance within an area. For example, the number of cases examined or positive cases identified by each service in an area. Ejiciency

(‘d)

Efficiency is a measure of the relationship between the outcome from a process and the input. Efficiency may be expressed as percentage efficiency (where inputs and outcomes are in the same units); as cost-benefit (where the benefits or outcomes are expressed in money terms and related to input costs); and as input/output ratios (where units are disporate). Cost/unit is extensively used in the study, e.g. cost/blood slide and cost/positive case. Procedures for determining performance are detailed in Tables 2 and 3. In some cases primary surveys were made of operational services. For other measurements, data was drawn from monitoring forms (records of dates and the outcome from activities).

DETERMINATION

OF COSTS

The goals of costing in the study were to determine the total cost of surveillance undertaken by each operational service and the elements of that cost; total cost to including, if appropriate, internal and external costs. An example is illustrated in Table 4. Expenditure (internal explicit costs) in the malaria service, as in all government organisations in Thailand is itemised under six major budget headings; personnel; remuneration; supplies and materials; public utilities; welfare expenditure; and buildings and fixed assets. The malaria service receives an annual budget with amounts assigned to each heading. The division allocates budgets to the five regions which in turn allocate budgets to zones. Zones provide operational services and occasional activities. The major issue in determining costs is how to apportion expenditure under budget headings to operational units at each level and finally to the operational services in the field. The system used is illustrated in Fig. 4. There are five stages to the system. Stage I:

Stage 2:

Stage 3:

apportionment of divisional, regional and zone budgets to divisional, regional and zone operational units respectively. apportionment of divisional operational units expenditure to cost of regional operational units and zone operational units. apportionment of regtonal operational unit expenditure to cost of zone operational units. assignment of items from operational units at division, region and zone to internal direct costs.

G.

HARDIXG

village

3. Malaria

volunteer

clinic

2. Malaria

Number

of +ve

Number

Number

Total

Number

Number

RT

radical

should treatment

by MVV

by MVV

cases undertaken

investigated

by MVV

by MVV

cases undertaken

in the zone

by MVV

within

diagnosis

as cumulative

time

of time and SD

for diagnosis

target

mean

Expressed

have been presented

for diagnosis -.__

in supplying

receive RP

cases detected

of patients

Total cost of malaria = ___.__

(MVV)

of fve

cost of malaria

3.3 Cost/investigation

=-

Total

ol cases detected

of cases detected

cases (MVV)

number

3.2 Cost/positive

zz

Number

(MVV)

received

should

of cases which

in providing

of cases which

3.1 Case detection

zz

Number

effectiveness

of fever cases which

clinic

number

2.7 Percentage

Total

=__

of malaria

of fever cases presented

effectiveness

clinic.

care. in each successive day, arithmetic

Number

2.6 Percentage

clinics

clinics

RT

PT

care at malaria

in malaria

clinics

by malaria

clinics

fever and seeking

seeking

having

clinics

in the zone

clinics

clinics

served by malaria

cost ol malaria

of fever patients

between

Population

2.5 Time

=-

Total

(MC)

investigated

cost of malaria

of patients

2.4 Cost/population

=

Total

cases-detected

(MC)

RT

have received

by malaria

should

cost of malaria

case (MC)

treatment

have received

PT

treatment

received

who received

radical

should

who

presumptive

of cases detected

Total

2.3 Cost/investigation

=

2.2 Cost/positive

who

cases detected

number

Number

Total

(MC)

ol people

2.1 Case detection

Number

of people

in providing

who

of people

of people

Number

1.7 Efiectiveness

Number

Number

1.6 EtTectiveness in providing

percentage

Cost/unit

Cost/unit

(%)

contribution

Relative

(%)

EtTectiveness

(2)

Effectiveness

Time

Cost/unit

Cost/unit

Cost/unit

(%)

contribution

Relative

(7;)

Electiveness

(2,)

Effectiveness

MS5

of costs from

source

MS5

Tuble

2

contirrued

overleaf

MS5

external

forms

and MS7

budgets,

and MS7

clinics

clinics

forms

costs from

at malaria

forms

and MS7

and MS7

at malaria

statistics

and MS5

and MS7

of 5Gl30 patients

source

survey

surveys and cases from

Analysis

Secondary

survey

and population

primary

forms

and MS5

MS7

survey

and

survey of 5CGl3 patients

budgets

budgets,

MSS

MSI

MSI

primary

source

source

source

budgets,

Secondary

Primary

Primary

From

From

From

Secondary

Secondary

Secondary

primary

forms

forms

1088

Sowm

KAEWSONTHI

and AL.A> G. HARDI%V(;

w .P

5

7

Division

Criteria Kl

*

Divisional operotlonoi units (71

. : * :

.

Regional budget

*2

Criteria

Region

%

Criteria c

*

Fig. 4. Apportion

;

(7)

units

: :

1 :

Regional operotionol

Regional operotionol units (7)

.

of costs.

KZ

Criteria

KZ

Criteria

Zone

e

)

units

16)

Zone operorionol

Zone operotlonal untts (6)

--c

r

Sub

zone

operation

3. Blood shde ehammation

Z Takmg blood >hdes

4ctivity

or wited

of house visitors

of target houses

(residenta at home etc.)

houses

questioned

Number of blood shdes taken

ei%ctiveness of taking blood slides (ACD)

of houses which should be wited

aclivilie,

Number

of caw

Number of blood blide examiniltwn?l

in target

I” a tmle

by muxoscopist

of m~croacoplst

Number of blood alides cxammed

elTxtwnra5

Total shdrh examined

3.6 Percentage

=

Total blrdrs ex,rmined - errors on chechmg

of identdication

whlcb should he reported

3.5 Pcrcentagc efiectnwws

to zone chief

~ourcz

MS8 forms

Secondary

results or blood ahde exammatwn

wtthln target ume

tn reportmg

Number of casx reported

rlfectnencss

3 4 Percentage

m target time

Secondary sourcs~- MS8 forms analysed serwce and occasional activity

blood shde examinatwn

3 3 Tome betwren takmg a blood sample and reporting results to LOW chief Expressed as cumulative percentage posItwe cases reported to LOW chief each successwe day after taking a blood sltde, arithmetic mean of time and SD

target time

target time

blood shdes and completmg

wthin

Number of cases which should be wthm

Number of cases examined

for each operational

for each oprratwnal

of house wsitors

of house visitors

of house visitors

Secondary source MS8 forms analysed serwce and occasional activity

survey-follow-up

survey--- follow-up

survey-hollow-up

source MSI forms

source MSI forms

3 3 Percentage etTectiveneas m collectmg

Primary

Primary

Primary

Secondary

Secondary

source of data

for each operatwnal

(?,I

Effectiveness

Etfectiveness (?“I

(“/,I

Effectiveness

EtTectivmess (%I

ElTectiveness (“,)

Type of criteria

of surveillance

Secondary source MS8 forms analysed service and occasronal actwity

to measure the performance

3 I Tame between taking a blood sample and completmg microscopic examination Expressed as cumulatwe percentage posmvr cases completed m each successive day after taking a blood sltde. arithmetrc mean of time and SD

Number of blood slides which should be taken

2 I Percentage

Number

of questioning

of households

efiectiveness

Number

I 4 Percentage

Number of houses vtsited on schedule ____.____ ~ ~ Number of houses which should be vwted

eflectiveness

Number

Number of houses responding

eKectneness

Number of target houses

1.3 Percentage

crtteria

en’ectneness of visttmg houses

Number of houses usned

Prrcentagr

1.2 Percentage

I.1

Performance

Table 3. Cnrur~a wxi

investigation

Case

7. Monitoring

of cases

6. Follow-up

5.

4. Treatment

blood

z

Number

Number

Number

Total

which

be made

form

form

which

of pages of each MS should

which

time

target -_.

time

be completed

of populot~on

1981

each day after taking

in 198%

treatment

proportions

are completed

in each column

column

to schedule

pages for each form

should

in each major _____-

of pages of each MS

Number

completed

of entries

of fully

number

7.2 Percentage

Number

= --_

up according in each column

of completed entries __-__

of entries

of cases to be followed

within

to schedule

IO schedule

up

up according

in follow-up

of cases

of cases followed

completion

Number

7.1 Percentage

Number

Number

effectiveness

number

time

target time --

cases identified

of follow-up

be treated

treatment

treatment

in target

treatment

treatment

radical radical

in target

of cases being followed ___

Total

Performance

6.2 Percentage

6.1

of positive

required

cases given radical

radical

and division)

of positive

of case investigation

of cases investigated

effectiveness

Number

Number

within

of cases to be cured

of cases cured

of radical

, .

slides from

be examined

and giving

should

with

ol cases which

effectiveness

region

mean of time and SD

of cases provided

Number

Percentage

5.1 Percentage

4.3

Number

sample

percentage

etTectiveness in providing

slides, arithmetic

4.2 Percentage

(zone,

should

of slides examined

a blood

Number

taking

slides which

activities

Hood

m each sector of nooulation

in taking

slide examination

as a cumulative

between

olx~ne

slides examined

of blood

of laboratory

Expressed

4.1 Time

Number

of blood

elfecliveness

per slide on blood

Cost = _-

3.8 Cost

Number

3 7 Percentage

(“a)

Performance

(“/b)

Performance

(:;J

Effectiveness

I%)

Performance

(%)

Ellectiveness

( %)

Effectiveness

( 7) 01

ElTectiveness

Time

Cost,lunit

(““J

Elfectivencss

MS5

All

MS

of villages

source--MS8

source-MS8

source-MS6

survey

source~~ MS8

forms

All MS forms

Secondary

Secondary

Secondary

Primary

Secondary

service and occasional

forms

Corms analysed

forms

and MS7

forms

forms

forms

forms

activity

MS5

source-MSR

and

source

analysts

Secondary

Cost

Secondary

for each opera&al

1092

SOMKID KAEWSONTHI

and ALAN G.HARDII‘;G

Table 4. Cost items of blood slide exammatmn: malaria diwslon Region

ZOIK

Salarles of mlcroscoplsts Laboratory supplies and materials Laboratory remuneration Public utilities for laboratory work

Direct explicit cost at reglonal headquarters

Direct enplut cost + at dwwonal headquarters

Depreciation of microscopes Depreciation of laboratory equipment

Direct Implicit cost at regional headquarters

Direct lmphclt cost + at dwslonal headquarters

Explicit overhead cost: administration cost and mamtammg of fixed assets

Indirect explicit cost at reglonal headquarters

IndIrect explut co,, + at divisIonal headquarters

Indirect implicit

Depreciation of buildmgs and fixed assets

Indirect implicit cost at reglonal headquarters

Indirect nnphclt cost + at dlwslonal headquarters

Direct

Direct supportwe cost: external funds direct to laboratory Travelling cost and other expenditure of patients in recewing blood examination

Direct explicit Direct impliclt Internal cost to division explicit

External cost to diwsion

Stage 5:

i

i

Direct implicit

{

Indirect exphcit

{

Indirect implicit

Stage 4:

Dlwslon

Time cost of patients Indirect supportive cost: external funds Indirect support to laboratory Expenditures of relatives attending patients Time cost of relatives and friends visiting patients

assignment of items from operational units at division, region and zone to internal indirect costs and apportionment of indirect costs to costs of surveillance. apportionment of internal direct costs and internal indirect costs to operational services and occasional activities.

Several criteria could be used for apportionment at each stage. For example, Divisional H.Q. budget is apportioned to seven operational units; administration, health education, applied research, chemical spray, entomology, epidemiology and laboratory. Some budget items such as chemical spray and drugs could be allocated directly to the user units. But expenditure such as salaries and materials has to be apportioned to operational units using suitable criteria. Criteria which, with some justification, could be used are equal distribution, the proportion of all types of staff in each operational unit or the monthly salaries of all types of staff in each operational unit. In the system developed a range of criteria can be applied at each stage of apportionment (Table 5). Since it is difficult to decide which is the most appropriate criteria to use for apportionment at each stage a system has been developed to allow study of the effects on using any combination of criteria in the three sets K,, K2 and K3 (Fig. 4). Using a network system and a computer programme data can be analysed to show the effect of different combinations of criteria on the cost of operational units and operational services. Maximum and minimum costs can be computed and the most reasonable cost determined by selecting a combination of the most suitable criteria. The system provides for direct costs of operational units at division, region and zone levels to be expressed as expenditure under each budget heading. Costs derived from divisional regional and zone budgets are subdivided into direct and indirect costs. Direct costs at this stage are expressed as expenditure under each budget heading but indirect costs appear as a lump

sum. Divisional, regional and zone expenditure is kept separate so that component costs can be identified at each stage (Fig. 4). In looking at the costs of operational services it is appropriate to view the Malaria Division as .the institution’. Budget expenditure on surveillance and monitoring activities at divisional headquarters. regional headquarters and within the zone are therefore internal costs. Costs external to the division are funds from abroad and costs of patients and relatives (Table 4). However it is possible using the system to readily identify direct expenditure on surveillance activities at sub zone level and the ‘overheads’ or external costs at zone, region and divisional levels. The network system allows apportionment of expenditure from budget headings. But one further issue remains, how to gather the data on which apportionment is based. Sources of data for apportioning costs in the malaria service are shown in Table 5.

m ANALYSIS OF SOME COSTS AND PERFORMANCES Although the primary aim of the research was to study how costs and performance of surveillance might be measured, it is also important to examine the data collected. Data was gathered in two zones by the research team and in a third zone by zone staff to field test the procedures. Some of the data gathered from two zones is presented in this section to illustrate some measures of performance, the cost of operational service, based upon vartous sets of criteria, and how data on costs and performance might be used by malaria managers.

The comprehensive monitoring system used by the Malaria Division (Fig. 3) provides. in principle, a valuable source of data on performance. Each case should be fully documented. In practice, an important

Cost and performance Table

5. Malaria

serwccs:

of malaria

types and source

surveillance

in Thailand

of crtter~a

for apport~onmg

Apportmnmg

1093 cost

of

Expenditure al division, region

and

zone to Type

proportion

cntena I. Personnel

based

1.1 Total

based

Units

of

staff

I.2 Civil

2. Salary

operational

Criteria:

of

servant

to zone and region to zone

Expenditure at zone to OperatIonal

Source

Records

on personnel. Division Dwision

Malaria

List of personnel

total

2.2 Monthly

cwil

salaries servant

at Zones Zones

interwews determme

wages

Report descriptmn

of&me

Chiefs

at Zones

under

survey and assistant

3. 6 and 7 to

the allocation

of personnel

to tasks.

2.4 Monthly temporary

Annual

3. 6 and 7 and Sectors

Chiefs

at

I

and job

3, 6 and 7. Primary

Zone

permanent

and salaries

and Region

employee

2.1 Monthly

for

criteria

Malaria employee

Temporary

of data

determining

services

I .4

salaries

based

at dwision

I.3 Permanent

2.3 Monthly

3. Responsibihry

Expenditure

wages Records

of Epidemiology

Section

3.2 Area

covered

Malaria

Division

I and

3.3 Real

expenditure

3. 6 and

7 Malaria

3. I Population

covered

Region

Division

at

Zones

Annual

Report

4. Functmnal

4. I Office 4.2 Field

based

personnel

4.4 Field

personnel personnel

of Interior survey

Zone

personnel

4.3 Office

Ministry Primary salaries salaries

by interview

chiefs and assistant

Zones

3, 6 and

determine

the allocation

to operational

Zone

chiefs of

7 are undertaken

to

of personnel

services at sub zone

level 4.5 Blood

slides examine

4.6 Positive 5. Time

based

5.1 Time

cases

allocated

to each task

and operational

services

Summary

Division

Primary

survey;

based

6. I Populatmn 6.2 Area 6.4 Time

at zone

their

daily

spent on each activity

index based index

based index

6.5 Equally 7.1 Direct amount

distributed allocation

of

to

specific/section/tasks 7.2

Report

index

6.3 Functional

7. Direct

Annual

all personnel

and sub zone levels record time

6 Index

of surveillance

Malaria

Direct

collection

Various

designs

of direct

cost data

proposed

collection

was undertaken

collection

but no direct

on

-

of cost for

each task operational SCIVCC

finding of the research was the low effectiveness of many monitoring procedures, Questions must therefore be raised about current monitoring practices. Measures of performance based on data drawn from monitoring forms must also be analysed with due consideration of the level of reliability and validity. Two examples of the performance of surveillance activities described in Tables 2 and 3 serve to illustrate how performance may be measured and expressed, conclusions drawn and issues raised. To reduce the probability of transmission the time between evidence of infection with malaria parasite and radical treatment should be as short as possible. The average time (days) between taking a blood slide and providing radical treatment, in 1980, for each operational service and occasional activity in two zones is shown in Table 6. The relative contribution ( “,) made by each service and occasional activity in the detection ofpositive malaria cases is also included. The effectiveness of operational services in providing radical treatment is presented in Table 7.

The average time for providing radical treatment, time between taking a blood slide and providing radical treatment to positive cases (Table 6) shows a similar trend in the two zones studied; malaria clinic, hospital, malaria village volunteer, village health centre, active case detection. While the mean time for providing treatment is marginally shorter in Zone A (4.0 days) than in Zone B (4.2 days) most of the operational services in Zone A take a statistically significant longer time than Zone B. This apparent contradiction occurs because a larger proportion of the cases in Zone A are detected at malaria clinics where the average time is short at only 1.3 days. But conclusions drawn from the data must be treated with caution. In both zones monitoring records (MS8 forms) are incomplete. There is no recorded date of treatment for 11.2‘A of cases providing blood slides in Zone A and 3.9% in Zone B. Study of hospital records also shows that a large number of malaria cases which are identified in hospitals are not recorded with the malaria service.

1094

KAEWSONTHI and ALANG. HARDING

SOMKID Table 6. Average

time between taking a blood sample

and providing

radical treatment Zone B

Zone A

Operational occasional

Relative contribution (%I

service and activities

1. Malaria clinic 2. Malaria village volunteer 3. Active case detection 4. Village health centre 5. Hospital 6. 7. 8. 9.

Case investigation Mass blood survey Follow up Special survey

Average time (Days)

Average time (Days)

1.3 11.1 15.9 13.8 8.8

51.7 17.4 16.6 4.2 3.4

1.2 7.8 8.1 8.4 5.9

s NS

0.5 0.1 0.1 1.4

6.9 7.1 23.8 2.4

0.1 2.0 0.0 4.6

10.0 6.1 0.0 6.0

NS NS s s

s

4.2 4.6

4.0 6.6 12,029

4387 cases

s s

71.6 14.1 4.7 4.9 2.1

Mean SD N Missing

Relative contribution (%)

t-test of slgmficant difference between means < 0.01

( %)

3.9

11.2

But even acknowledging doubts about the reliability providing immediate testing and treatment compared of data the longer average time taken to provide with the slow turn round time of active case detection is treatment by operational services in Zone A coupled also clearly illustrated. with the high percentage of incomplete records in the Any relationship between the number of active same zone suggests that Zone A has less stringent carriers in a zone and the effectiveness in providing control of surveillance and monitoring operations diagnosis and radical treatment is subject to three than Zone B. factors in addition to the reliability of data. The average time for radical treatment by each (i) many cases could and will be active carriers operational service (Table 6) provides a useful target to before they present themselves for a blood test. In a be improved by each zone and a comparative study of malaria clinics 20 “/‘,of patients took 10 days measurement among zones. But average time cannot before seeking treatment after the onset of symptoms. show the proportion or number of patients treated by (ii) there is some doubt about the time period during each operational service who may become active presumptive will prevent which treatment carriers as the temporary curative effect of presumptive transmission. treatment decreases. The latter can be more clearly (iii) radical treatment may not effect a cure and considered when performance in providing radical prevent transmission. Systematic follow up of a small treatment is expressed as percentage effectiveness number of cases in the two zones shows that 20 p, were (Table 7). Percentage effectiveness in providing radical still positive after 4 months. These cases may be viewed treatment I . n number of cases treated withm a target ttme = number of cases which should be treated within the target time’ I If it is assumed that patients become active carriers IO days after receiving presumptive treatment, 10 “/: of all patients will become active carriers for some days. With IO days as the target time Zone A shows a consistently poorer performance than Zone B for all operational services. The value of malaria clinics in

Table

Effectiveness

7.

Effectweness

of operatlonal

( ;;,I

I 5 I0 15 20 Number Percentage

etrectlveneas

services in providing Malaria village volunteer zone

MalaW clinic Zone

All srrvices itobnDksh

as reinfected but could equally well be non-responsive to drug therapy or could have failed to take the drugs prescribed. The simple measure of effectiveness in providing radical treatment is a useful tool for evaluating the performance of operational services. But it also raises

FlA

60.6

96.9

97.I

70.5 89.2 96.2 98.4

98.0 98.7 99.4 99.7

9X.6 99.4 99.9 100

of cases treated wthin

of cases which should be treated

Actwe cilse detectlon Zone 014

10.4 26.4 54.3 74.7 90.6

IO.2 29.8 7X.2 95.0 98.6

target time

= Number

radical treatment

wthm

target time

Village health centre

B/A 2.4 7.x 27.5 60.5 75.4

Hospital Zone

20X

1.X 26.9 79.9 95 3 98. I

BlA 137 21.4 37.9 61.0 7x.0

71 30.6 76.0 91.6 97.0

B 41.0 53.x 69.2 74.4 x7.2

21.1 46.7 xx.4 97.3 99 7

Cost and performance

of malaria

Costs Analysis of data on the costs of surveillance in two zones is not completed. But at this stage preliminary analysis can be made of the costs of operational units at division, region and zone levels following apportionment of expenditure under the standard budget headings. Percentage costs of operational units at division, Region X and Zones A and B using one set of criteria for apportionment are shown in Table 8. Criteria used in apportionment (alternative 6) are shown in Table 9. If the criteria used are accepted as a reasonable basis for apportionment three conclusions may be drawn. (i) Division L’Sregion. The division and Region X show a similar pattern of costs; approx. 40 “/I of total expenditure on administration, approx. 3 y0 for research and approx. IO:/, for each of the other operational units. It is perhaps surprising that the Division and region should show a similar pattern of expenditure given their different functions.

region and zone expenditure

apportioned

to operational Operational

1 xt

(

1 Diwsion Reglon X Zone A

1ZoneB

H.Ed

1

Insect.

1

Lab.

1

Epid.

1

42.9

(

12.0

1

11.4

(

12.8

1

9.0

(

II.4

8.5

1

18.5

1

2.8

[

25.9

I

9.6

1

2.4

1

42.1

for apportionment

Budget item

1095

Criteria

14.5

6.1

for apportionment

2. 2.1 2.2 2.3 2.4 3.5 3.6

Remuneranon Fees Freight Travelling allowance Transportatron Mamtenance of fixed assets Others

i ;:I 3.2 3.3 3.4 3.5 3.6 -3.7 3.8

Supplies and materials Office supplies Drugs lnsecticldes Vehicle supplies and fuel Maintenance of assets Laboratory supplies Health Education supphes Others

Direct to laboratory Direct to health education Total number of personnel

4.

Public utihties

Total number

8.5

1 (

II.6

6

3.1

1 Treat. I

-

-

in each operational

umt)

I

alternative

-

19.1

6

wage of staff in each “nit wage I

Total number of personnel Direct to chemical spray Total monthly Total number Total staff

salaries of staff of personnel

Equally distributed Excluded at Division allocated to treatment Equally

and Region since bulk purchased and insecticides.

distributed

of personnel

Equally distributed Proportion of low income staff in each umt

I

23.3

I

20.6

(proportion

Res.

2.2

27.3

1

of budget expenditure:

Proprtlon of salaries Proportion of permanent Proportion of temporary ProportIon of salaries

Ent.

7.3

2.2

1

region and zone: alternative

umts at each level

1

Table 9. Criteria

i Welfare expenditure ;:I Social welfare 5.’ Low mcome support

units at division,

Admm.

444

Salaries and subsistence Permanent wage Temporary employees Allowances

in Thailand

(ii) Diuision/Region X us zones. The distribution of costs to operational units at zone level is significantly different to that at division and Region X. The proportion spent on administration, health education and laboratory is lower, while epidemiology and entomology are similar. A higher proportion of costs at the zone level are spent on insecticides and spraying with approx. 20 7: of zone budgets spent on treatment. Lower administrative costs at zone level, expenditure on treatment and a large proportion of budget to spraying and insecticides is to be expected in view of zone operations. The high cost of laboratory activities at Region X and division, presumably due to training of personnel and re-checking of slides, warrants more detailed study. (iii) Zone us zone. Zones A and B cover roughly the same area. Zone B is mountainous with a sparce population and a high incidence of malaria. A larger proportion of Zone B’s budget is therefore spent on insecticides and spraying. The significant difference in the proportion of budget spent on administration in the two zones is surprising. But judgement can only be made in relation to the budgets, work load and effectiveness of operational services in the two zones.

about monitoring procedures, transmission and treatment which are addressed in the section ‘Use of data by malaria managers’. questions

Table 8. Divwon.

surveillance

for zones. At zone level dxectly

1

1096

SOMKID KAEWSONTHI and ALAN G. HARDING

Table 10 shows a more detailed breakdown of antivector and antiparasite costs for the two zones. (Costs are internal costs to the zone without, at this stage. consideration of the’overheads’ or external costs from division and Region X.) Higher expenditure on insecticides and spraying in Zone B accounts for most of the difference in the budgets of the two zones. Zone A would therefore appear to be less efficient in its administration, spending 0.8 x IO6 Baht on administering a budget of 4.5 x IO6 Baht while performing less spraying, taking a smaller number of blood slides than Zone B and having a poorer performance in terms of the time between taking a blood slide and providing radical treatment. Comparison of the cost of antiparasite measures in the two zones reveals some effects which may not be fully explained by differences in terrain, population. incidence of malaria and relative contribution of operational services (Table 6). Is the higher laboratory costs in Zone B due to less use of malaria clinics, to the existence of more units or other causes’? Is the cost of epidemiology independent of the number of blood slides taken and number of positive cases‘? Both zones spent the same amount on epidemiology. Is the lower cost per case in Zone B solely due to the higher incidence of cases. If this is so, it might be more informative to express performance as the cost/case prevented rather than the cost/case.

C’se of data hy malaria managers How could or will malaria managers make use of the procedures developed and information on performance and costs’? The simple answer to the question is that we do not know. Although there has been very strong support for the research from the Director and other staff within the region and zones examined, it has to be acknowledged that other staff within the malaria service may be sensitive to attempts to evaluate performance and to compare costs of operations. The extent to which the procedure and data are used will therefore probably depend on four factors: (i) the skill and political sensitivity with which the procedures are introduced and the results of this particular study communicated to staff within the Malaria Division: (ii) the relevance of costs and performance measurements to the control and improvement of services. (The measurements must have meaning to personnel and relate to day to day operations at each level of the malaria service); (iii) feasibtlity of gathering data by zone and regional personnel (after training) with the resources available: (iv) reliability of data (low reliability will invalidate conclusions and disillusion those who give their time to the collection and analysis of data).

Assuming statf in the Malaria positive attitude. we see five major

Division

take

a

uses for data on

costs and performance.

(i) Analysis of how money is spent at each level in the Malaria Division; cxpcnditure on operational units and services in relation to the goals of the malaria xrvicc.

(ii) Establishment of performance and cost targets to be met and improved. (iii) Evaluation of the relative contribution and cost/unit of various types of service, and the economic merit in encouraging particular types of operational services. (iv) Determination of the best combination and distribution of operational services for maximum effectiveness and minimum cost. (v) Comparison of the costs and performance of different zones with due recognition to geographical conditions and population distribution. Uncertainties will remain about application of the procedures developed. But the research has highlighted several issues of importance in efforts to improve the malaria control programme. Many of the issues relate to the effectiveness of surveillance and monitoring procedures, developed long ago for an eradication programme, which are now used in a control strategy. For example: (i) Analysis of monitoring in two zones in 1980-198 I disclosed a very low effectiveness to many monitoring procedures. Unless the effectiveness of monitoring is improved and/or the approach to monitoring modified, malaria control will be less effective and doubts will exist when measurements are made of the performance of surveillance activities. (ii) Follow up of positive cases is required at I, 23.6. 9 and 12 monthly intervals. The effectiveness of this procedure was found to be about 4”;. Since the number of patients cleared of infection is not known there could be an increase in the number of carriers and a distortion of statistics as patients not cleared of infection are registered as new cases. (iii) Blood samples are only taken from persons reporting a fever. In a study of 4659 persons in three villages, 1.02 “/‘,of the population was found, on blood slide examination. to be positive. Only 209; of these cases reported having a fever. The high proportion of symptomatic cases raises questions about the relevance of established procedures.

HEALTH

ECONOMICS RESEARCH COUNTRIES

IN DEVELOPING

The second aim of the research was to strengthen the research capability in health economics at Chulalongkorn University, Bangkok, Thailand. The educational outcome is, in many respects as important as the procedures developed and data collected. The UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases which funded the projects seeks to strengthen selected centres in developing countries so that staff within the centres may carry out both research and training. TDR argue that such activities “will help to provide the manpower needed to enable the tropical countries to implement their own policies and programmes for research on disease control” (WHO undated). The effects of this project on the competence and confidence of the research team are fully reviewed elsewhere (final report). What is more pertinent at this point in time is to review the lessons learned and consider how health economics research on disease

Cost

and performance

of malaria

surveillance

in Thailand

1097

Table IO. Costs of antivector and antiparasite measures

Note: Zone A

Population X IOh Blood slides x IO4 Posmve slides x IO”

control programmes

1.3 107 0.7

Zone B 0.5 132 1.2

could be better stimulated and supported in developing countries. The TDR philosophy of supporting the development of research teams in countries where the six major tropical diseases are endemic is to be applauded. But some improvement in the response to WHO initiatives and in the quality of research completed might be expected if training and professional support was given to potential and practising-researchers. (i) Training programmes should be provided at universities in developing countries for potential and practising researchers. The short training programmes should provide experience in the submission of research proposals and the management of research projects through the use of simulations and case studies. (ii) A system of advisors should be established to assist new principal investigators in the planning and management of research. Advisors should have extensive experience in research but not in the actual topic of the research study. Through such a system the principal investigator would retain responsibility for the direction and progress of the project while drawing

on the experience

of the advisor. Both advisor and the principal investigator would probably benefit from the cross disciplinary interaction. The reason for these proposals is the lack of experience of research in many developing countries. Although a number of academic staff in universities in a country such as Thailand may have higher degrees,

there are seldom any established schools of research (particularly in health economics). Since those with higher degrees have frequently responded to initiatives rather than managed their own research programmes there is no pool of staff experienced in research management and few, given cultural conditions, with the competence or confidence to initiate research proposals. The result is a chicken and egg situation in which lack ofexperience makes it difficult for potential researchers to initiate and successfully complete research in health economics.

Acknowledgement-The investigation received support from the UNDP/World Bank( WHO Special Programme for Research and Training in Tropical Diseases.