Cost of Delivering Child Health Care Through Community ... - HSRIIhttps://www.researchgate.net/...Health...Health.../Cost-of-Delivering-Child-Health-Care...

3 downloads 0 Views 363KB Size Report
accredited social health activists (ASHA) from intervention and control areas to collect cost data using an economic perspective. Bootstrap method was used to ...
Journal of Tropical Pediatrics Advance Access published July 19, 2013

JOURNAL OF TROPICAL PEDIATRICS, 2013

Cost of Delivering Child Health Care Through Community Level Health Workers: How Much Extra Does IMNCI Program Cost? by Shankar Prinja,1 Sarmila Mazumder,2 Sunita Taneja,2 Pankaj Bahuguna,1 Nita Bhandari,2 Pavitra Mohan,3 Henri Hombergh,3 and Rajesh Kumar1 1 Post Graduate Institute of Medical Education and Research, School of Public Health 2 Centre for Health Research and Development, Society for Applied Studies 3 UNICEF India Country Office Correspondence: Shankar Prinja, Assistant Professor of Health Economics, School of Public Health, Post Graduate Institute of Medical Education and Research, Chandigarh, India. Tel: 91-9872871978. E-mail .

Background and methods: In the setting of a cluster randomized study to assess impact of the Integrated Management of Neonatal and Childhood Illnesses (IMNCI) program in the district of Faridabad in India, we randomly selected auxiliary nurse midwives (ANM), anganwadi workers (AWW) and accredited social health activists (ASHA) from intervention and control areas to collect cost data using an economic perspective. Bootstrap method was used to estimate 95% confidence interval. Results: The annual per-child cost of providing health services through an ANM, AWW and ASHA is INR 348 (USD 7.7), INR 588 (USD 13.1) and INR 87 (USD 1.9), respectively. The annual per-child incremental cost of delivering IMNCI is INR 124.8 (USD 2.77), INR 26 (USD 0.6) and INR 31 (USD 0.7) at the ANM, AWW and ASHA level, respectively. Conclusion: Implementation of IMNCI imposes additional costs to the health system. A comprehensive economic evaluation of the IMNCI is imperative to estimate the net cost implications in India. Key words: cost, incremental cost, child health, IMNCI, health worker.

Introduction The Integrated Management of Neonatal and Childhood Illnesses (IMNCI) program is implemented in 433 districts of India and aims to prevent childhood

Funding The study was funded by the World Health Organization, Geneva (through an umbrella grant from USAID); the United Nations Children’s Fund, New Delhi; and the GLOBVAC Program of the Research Council of Norway through grant No 183 722. Individual scientists at WHO and UNICEF contributed importantly to the planning, analysis and reporting of this study. However, the central bodies of these agencies and the Research Council of Norway had no influence on how the data were collected, analyzed or presented. The corresponding author had full access to all data that were analyzed and had final responsibility for the decision to submit the manuscript for publication.

mortality through actions at the household and health system level. At the household level, general health care-seeking behavior of families is improved to enhance the uptake of preventive services and promote good child-rearing practices. From a health system perspective, health care providers are trained to improve their skills for diagnosis and management of childhood illnesses. Finally, the health system is strengthened by improving the supply of drugs at the primary care level and by improved referral pathways [1]. Several studies evaluating Integrated Management of Childhood Illnesses (IMCI) implementation costs and effectiveness from a societal perspective have noted either cost-neutral or cost-saving effects on national health care expenditures [2–5]. However, using a health system perspective, another study from Bangladesh found that implementation of the IMCI program would result in an additional cost of approximately 2.6–4.0 million US dollars (USD) on account of extra health workforce, about 1–1.5% of the total health sector budget of the Government of Bangladesh [6, 7].

ß The Author [2013]. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected] doi:10.1093/tropej/fmt057

1 of 7

Downloaded from http://tropej.oxfordjournals.org/ by guest on July 20, 2013

Summary

S. PRINJA ET AL.

Methods We conducted our study in district Faridabad of Haryana state, which has a total and under-five population of 2.7 million and 0.39 million respectively [9]. The IMNCI program was implemented in nine cluster-randomized primary health centers (PHCs) of the district, whereas routine child health care services were implemented by nine other PHCs [8]. Detailed description of the intervention is available elsewhere [8]. We selected 8 and 10 ANMs and ASHAs from intervention and control areas, respectively, whereas 9 AWCs were selected randomly from intervention and control areas for costing of child health services at the anganwadi level. Cost data collection and analysis We used a bottom-up ingredients approach to estimate the economic cost of child health care services delivered by the primary care workers, with and without IMNCI. Trained field investigators, with graduate-level qualification, collected the data on resources, i.e. building space, equipments, salary and any incentives or allowance for human resources, medicines, vaccines, other consumables and diagnostics for 1 year, i.e. 2010. Data were collected from the AWW, ANM and ASHA workers (Table 1). All the resources consumed at each level were apportioned for child health care using a specific methodology (Table 1). The equation mentioned in the following text describes the estimation of time cost of primary care workers for under-5 child health. U5

CijðsalaryÞ ¼ fSijðannualÞ  1=n  fðU5 Tij U5 Cij Þ þ ½ðn  ððU5 Tij U5 Cij Þ þ ðA5 Tij A5 Cij ÞÞ U5 Pij gg

2 of 7

where U5Cij health

(salary) ¼ Annual

salary for under-5 child

Sij(annual) ¼ Annual salary U5 Tij ¼ Average time taken to treat per child (under-5) U5 Cij ¼ Average under-5 child contacts per day A5 Tij ¼ Average time taken to treat per above-5 individual A5 Cij ¼ Average above-5 child contacts per day U5 Pij ¼ Total under-5 population/Total service population ¼ Proportion of non-care provision time devoted to child health (home visits, attending meetings, maintaining records, writing reports, etc.) n ¼ Number of working hours in a day   i ¼ 1 ðANMÞ, 2ðASHAÞ, 3ðAWWÞ j ¼ 1 ðInterventionÞ, 2ðControlÞ These ratios (U5Pij) were also used to apportion the child health resources in terms of drugs, consumables, space, equipments and diagnostics that were used jointly for both under-5 and above-5 population (Table 1). Drugs that were consumed solely for child health contributed totally, whereas those that were used solely for adult health were not accounted for in child health care cost estimation. Capital equipments, i.e. those that yielded returns lasting >1 year, were annualized based on the useful life of the capital item and by assuming a discount rate of 3% to account for time preference of money [10]. For computing the cost of space, we estimated the area in which the sub-center or anganwadi center was built and used the currently prevailing rental price for land in the village to estimate the opportunity cost of the infrastructure. Resources spent on training the community-level workers were also estimated in terms of time costs of trainees and trainers, space, equipments, training material and travel costs. Fraction of these resources attributable to child health was estimated based on the content of the training. Those trainings (e.g. family planning training) that did not have any child health component were excluded from costing. Those trainings that were solely aimed at improving child health care (e.g. neonatal resuscitation) were attributed entirely for the same. Trainings such as skilled birth attendance, management information system, etc, which aimed at improving both child and adult health, were apportioned by multiplying the same with U5Pij. Based on discussion with local health care managers, we assumed that the refresher trainings will not be held before the seventh year; hence, we estimated the annualized economic cost of training using a discount rate of 3%. The cost of supervision and program support for delivery of child health services by CHWs was estimated at the district and sub-district level. The time cost of staff for supervision was apportioned by the Journal of Tropical Pediatrics

Downloaded from http://tropej.oxfordjournals.org/ by guest on July 20, 2013

In India, several factors make implementation of this child health program unique. First, IMCI was adapted with a strong focus on the neonatal component and hence renamed as IMNCI. Second, the IMNCI program had a greater emphasis on the community component, in terms of behavior change communication for improved child care practices. A cluster randomized trial from North India found significant reductions in neonatal and infant mortality [8]. We designed an economic evaluation that was nested in this IMNCI effectiveness trial. In this article, we report the cost of delivery of child health care services through community health care workers’ (CHWs) level. For our analysis, we include auxiliary nurse midwives (ANM), anganwadi workers (AWW) and accredited social health activists (ASHA). Third, we estimate incremental cost of child health services at each of these levels with implementation of IMNCI program, as compared with delivery of routine child health services without IMNCI.

Pay slips and tour diary of staff involved in supervision Haryana Government procurement prices Log book of vehicles for distance, rate contract of Haryana Government for vehicle hiring Petrol bills, and travel allowance (CS, PI records) Consumable register (CS, PI records)

Proportion time spent for monitoring

Proportion distance travelled for child health monitoring

Proportion time spent for monitoring Proportion distance travelled for child health monitoring

Proportion time spent for monitoring

Proportion of time spent on under-5 child health care Proportion of time spent on under-5 child health care Proportion of time spent on under-5 child health care Proportion of time spent on under-5 child health care Proportion of time spent on under-5 child health care Only for intervention area, and for under-5 care, hence full allocation Proportion of under-5 children among total population benefiting from IEC Apportioned between ASHA and AWW according to the proportion of meetings SBA training: Proportion of live births to live births plus pregnant womenASHA training: Proportion of under-5 population among under-5 plus 15–45 years female population

PCW: primary care worker; CS: civil surgeon; PI: agency implementing IMNCI in the intervention area of district; ASHA: accredited social health activist; AWW: anganwadi worker.

Consumables

Travel (Monitoring)

Costs at higher levels Salary (district and sub-district staff involved in monitoring) Equipment (i.e. computers) Vehicles

CS, PI office records

Social mobilization costs (women meetings) Training (honorarium, travel, space, equipment and time costs) CS, PI office records

CS, PI office records

Facility survey by study investigators Non-consumable register (PCW) CS, PI office records Consumable register (PCW) Salary slip (PCW) CS, PI office records

Costs at the PCW level Building (space) Equipment Travel Drugs and consumables Salary Incentives

Allocation statistic for joint costs

Downloaded from http://tropej.oxfordjournals.org/ by guest on July 20, 2013

Journal of Tropical Pediatrics

IEC costs (banners, wall paintings, etc)

Data source

List of items

TABLE 1 Data sources, assumptions and allocation statistics used for analysis

S. PRINJA ET AL.

3 of 7

S. PRINJA ET AL.

proportion of man-hours that the personnel spent travelling for child health supervisory activities during the past 1 month. Tour diary of the personnel was reviewed to assess the same. This factor was also used to apportion the other capital and recurrent costs of supervision.

Results The average population served by an ANM in the study area was 5500, whereas an AWW and an ASHA were providing services to about 1500 and 1700 individuals, respectively (Web-appendix Table A1). Each ANM, AWW and ASHA in the areas with standard health care delivery programs attended to 3.1, 2.3 and 4 sick under-5 children per day (Webappendix Table A2). In contrast, an AWW in an IMNCI implementation area catered to 3.3 sick children per day, whereas both ANMs and ASHAs managed 7 sick children per day. We found that the average time spent by an AWW and ANM in treating a sick under-5 child was 11.4 minutes. Similarly, ASHA took 13.2 min to examine and treat a sick child (Web-appendix Table A2). With IMNCI, the health workers took longer time to manage sick children using the algorithms. On an average, ANM, AWW and ASHA took 21.4, 52.6 and 4.5% extra time, respectively, to manage a sick child than their counterparts who were not trained in IMNCI. Cost of routine child health services in the control area The overall annual cost of child health care was INR 337 050 (USD 7490), INR 103 809 (USD 2307) and INR 15 859 (USD 352) per ANM, AWW and ASHA level, respectively (Table 2). Major drivers of child health care cost at the AWC and ANM level were capital and human resources (Fig. 1). Human 4 of 7

Incremental costs of IMNCI at the health worker level With IMNCI, child health services at the ANM, AWW, ASHA level costs the health system INR 472 (USD 10.5), INR 614 (USD 13.6) and INR 170 (USD 3.8) per child, respectively. We found an overall increase of INR 125 (115–134), INR 26 (14–39) and INR 83 (81–84) in per-child cost of health care services at the ANM, AWW and ASHA level, respectively, when implementing IMNCI (Table 3). The incremental costs at the ANM level were contributed primarily by the additional time spent on child health (54.7%, INR 68 per child per year), monitoring and supervision (20.9%, INR 26 per child per year) and better availability of drugs (8.8%, INR 11 per child per year). At the ASHA level, incremental costs are mainly attributed to monitoring and supervision (64.4%, INR 53 per child per year), increased time contribution of ASHA (19.1%, INR 16 per child per year) and augmented drug supply (16.4%, INR 14 per child per year). Discussion Overall, we found that the cost of providing health care services to a child through an ANM, AWW and ASHA costs the Government of India an average of INR 348 (USD 7.7), INR 588 (USD 13.1) and INR 87 (USD 1.9), respectively. Providing child health care services, with IMNCI, additionally cost INR 125 (USD 2.8), INR 26 (USD 0.6) and INR 83 (USD 1.8) at the ANM, AWW and ASHA level respectively. Whereas major drivers of cost at the ANM level are staff time, drugs and capital cost, staff time and drugs contribute the major expenses while providing services through AWW. Incremental costs of implementing IMNCI at the primary care worker level are mainly contingent on the additional time allocated to child health, program monitoring and strengthening of infrastructure in terms of increased availability of drugs. To our knowledge, this is the first study from India that computes the cost of comprehensive child health (and nutrition for AWW) care services at the primary care community health worker level. Implementation of the IMNCI program in a research mode with a cluster randomized study design in a district provided a unique opportunity to study the incremental costs of the program at the community health care worker Journal of Tropical Pediatrics

Downloaded from http://tropej.oxfordjournals.org/ by guest on July 20, 2013

Statistical analysis All costs were converted to 2010 prices (Indian National Rupees, INR), with cost in USD estimated at a conversion price of INR 45 per USD. Overall cost of providing child health services was estimated by summating the cost incurred for child health on account of time of personnel, space, equipments, medicines, vaccines, consumables, training and cost of monitoring and supervision. Per-unit cost of service delivery was obtained by dividing the overall cost of child health care services by the number of under-5 children enrolled for care provision by the primary care worker. To account for a small sample of health workers selected for estimation of cost, we used the bootstrap method to calculate the mean cost and its 95% confidence limits. Statistical significance of difference of mean costs in intervention and control areas was estimated using the t-test, and 95% confidence interval for the mean difference was computed.

resources (56.7%) and drugs (32.4%) were significant cost drivers at the ASHA level. Costs for monitoring and program support constituted 1.6 and 10.9% of total child health care costs at the ANM and ASHA level, respectively. Cost of health services per child at the ANM, AWW and ASHA level was INR 348 (USD 7.6), INR 588 (USD 13.1) and INR 87 (USD 1.7), respectively (Table 3).

S. PRINJA ET AL.

TABLE 2 Annual cost of child health services at the primary health care worker level, with and without IMNCI Primary care worker

23 335 760 1128 79 097 36 103 809

Standard services plus IMNCI INR (USD)

Percent change

(518.6) (16.9) (25.1) (1757.7) (0.8) (2306.9)

27 941 3602 3460 73 514 1778 109 670

(620.9) (80.0) (76.9) (1633.6) (39.5) (2437.1)

19.7 373.9 206.7 7.1 4838.9 5.6

86 016 (1911.5) 58 896 (1308.8) 4009(89.1) 167 393 (3719.8) 2375 (52.8) 5110 (113.6) 337 050 (7490.0)

151 241 68 649 1947 186 910 5874 24 701 461 056

(3360.9) (1525.5) (43.3) (4153.6) (130.5) (548.9) (10245.7)

75.8 16.6 51.4 11.7 147.3 383.4 36.8

8964 (199.2) 5235 (1116.3) 0 (0.0) NA 32 (0.7) 1703 (37.8) 15 859 (352.4)

12 049 (267.8) 7799 (173.3) 4 (0.1) NA 2503 (55.6) 8918 (198.2) 31 257(694.6)

34.4 49.0 0 0 7721.9 423.7 97.1

ANM: auxiliary nurse midwife; ASHA: accredited social health activist; AWW: anganwadi worker; INR: Indian National Rupees, USD: US Dollar.

level without introducing the biases of a case–control study. The additional time required to manage a single child by the AWW and ANM in our study is similar to that reported from Bangladesh [7]. Major determinants for cost of child health care services at the ANM level are capital and staff time. Similar observations have been found in Tanzania and Brazil [5, 11]. Another Indian review estimated the cost of child care services at the anganwadi level to be USD 10, which is close to our estimate [12]. High number of cases treated by ANM, ASHA and AWW per day in areas implementing IMNCI could be because of higher morbidity in IMNCI areas. However, a systematic review of the impact of the IMNCI intervention shows a decline in childhood morbidity rate in a wide range of settings where IMNCI was implemented [4], hence we believe that the rise in treatment load was an evidence of shift in treatment-seeking behavior from private practitioners or local quacks to the primary care community health workers. Moreover, findings from India also indicate that IMNCI results in lowering of neonatal and infant mortality [8]. We note the limitations in our study methodology and generalizability of findings. Due to constraint in resources, we did not undertake rigorous time Journal of Tropical Pediatrics

allocation study, and rather relied on reported time spent by study participants per child contact, and supplementing the same with reported number of under-5 child contacts as reported in service registers. We could not collect data on cost of supervision for AWW, as the same was being done by a department other than the health department, hence our results for AWW should be viewed in light of this limitation. Finally, the study covered one district from North India. Considering the vast diversity in health care delivery among Indian states, we recognize that there could be variations to the base estimate of the unit cost. The additional cost of delivery of IMNCI has fiscal implications for its implementation. However, full economic evaluations of the IMNCI program that take a societal perspective have shown that the reduction of morbidities and the shift in the treatment seeking from secondary care to primary care health workers result in reduction in the treatment costs, which offsets the additional investments made as part of the program implementation. Multi-country evaluations from Brazil, Tanzania and Uganda point in this direction. We believe that with higher out-ofpocket spending levels at private providers in India, reduction in treatment costs can result in an even greater cost-saving impact of the IMNCI program. This calls for a full economic evaluation of the 5 of 7

Downloaded from http://tropej.oxfordjournals.org/ by guest on July 20, 2013

AWW Human resource Drugs Consumables Capital Training Total ANM Human resource Drugs Consumables Capital Training Program support Total ASHA Human resource Drugs Consumables Capital Training Program support Total

Standard child care services INR (USD)

S. PRINJA ET AL.

TABLE 3 Mean annual per under-5 child cost of primary health care services in the district of Faridabad, Haryana, with and without IMNCI program Primary care worker AWW ANM ASHA

Setting

No IMNCI IMNCI No IMNCI IMNCI No IMNCI IMNCI

Mean per-child cost [INR(USD)] 588 614 348 472 87 170

(13.1) (13.6) (7.7) (10.5) (1.9) (3.8)

SE mean

4.6 4.4 2.3 4.3 0.3 0.5

Mean difference [INR (USD)]

p-value

26.1 (0.6)

95% CI for mean difference LL

UL