Efficiency in Healthcare Sector in Assam: A Sub-State Level Analysis

Purohit BC

ISSN 0972-5997

Published Quarterly

Mangalore, India

editor.ojhas@gmail.com

Custom Search

OJHAS Vol. 14, Issue 4: (October-December 2015)

Original Article
Efficiency in Healthcare Sector in Assam: A Sub-State Level Analysis

Author
Brijesh C Purohit, Professor, Madras School of Economics, Kottur, Chennai - 600025

Address for Correspondence
Brijesh C Purohit,
Professor,
Madras School of Economics,
Kottur, Chennai - 600025,
India.
E-mail: brijeshpurohit@gmail.com

Citation
Purohit BC. Efficiency in Healthcare Sector in Assam: A Sub-State Level Analysis. Online J Health Allied Scs. 2015;14(4):1. Available at URL: http://www.ojhas.org/issue56/2015-4-1.html

Open Access Archives
http://cogprints.org/view/subjects/OJHAS.html
http://openmed.nic.in/view/subjects/ojhas.html

Submitted: Nov 13, 2015; Accepted: Dec 26, 2015; Published: Jan 30, 2016

This work is licensed under a Creative Commons Attribution-No Derivative Works 2.5 India License

This Article

Abstract: This paper aims at exploring efficiency in health care sector at sub-state level using district level information for the years 2010-11 and 2011-12. We estimate a potential level of efficiency that could be achieved by a district health system. A panel data set is used comprising of 23 districts of Assam. Utilizing a stochastic frontier model our results indicate that better availability of medical manpower including doctors, nurses and midwives has led to an optimum utilization of existing medical institutions and beds. Thus based on achievement of their existing normative potential, districts like Dhemji, Kamrup and Barpeta are three top ranking districts. The low performing districts based on this criteria, including Darrang, Dhubri and Kokrajhar have been not able to utilize existing medical institutions and beds capacity due to the constraint of inadequate medical manpower, higher population density, higher rural populations, lower literacy levels and lack of comparable roads development relative to efficient districts in the state. Thus the overall efficiency in the district level health system could be improved to bring down infant mortality in lower performing district with more focus of policy on the factors that have been highlighted by our study.
Key Words: Healthcare sector, Rural areas, Sub-state level analysis.

Context:

Emphasis on efficiency in resource utilization in healthcare sector is now considered to be a trend in resource allocation across social sectors. This is witnessed in literature by a number of studies which have laid emphasis on the overall health system performance and its impact on health outcomes.[1,2] Some country specific studies, for instance, have concentrated on hospitals, nursing homes, HMOs and district health authorities.[3-9] As methods both parametric or non-parametric techniques have been employed. Among the later methods, an idealized yardstick is developed which is used to evaluate economic performance of health system. More frequently, due to its normative approach, a frontier efficiency measurement technique is used. It provides a production possibility frontier depicting a locus of potentially technical efficient output combination that an organization or health system is capable of producing at a point of time. An output combination below this frontier is termed as technically inefficient.[10-12]

Indeed there exists an exhaustive review of studies in health care sector which provides us in detail the steps and empirical problems that have been highlighted by researchers.[3,4,13] There are also some studies in the developing countries’ context and particularly in the Indian context that mostly focussed either on all-India rural or urban sector or the analysis has been carried out up to state level aggregates only. Yet, a limited number of studies have also focussed on District level analysis, which pertain to states including Punjab, Maharashtra, Karnataka, West Bengal and Madhya Pradesh.[14] We extend our analysis in this paper to focus on efficiency of health care system at sub-state level (i.e., district level) in India using Assam state and its district level data. We explore the reasons for relative performance of different districts with frontier estimation technique.

Assam is a north eastern state of India and comprises an area of 30,285 square miles. Assam is surrounded by seven other Indian states namely: Arunachal Pradesh, Nagaland, Manipur, Mizoram, Tripura, Meghalaya and Sikkim. As per 2011 census, total population of Assam was 31,169,272 and it has increased (from 26,638,407) in the last ten years with a growth rate of 16.93%. At present with a number of new districts added recently, there are 32 districts in the state.

A comparative view of Assam and India is presented in Table 1 below. It can be observed that the life expectancy of Assam both for male and female at 61 and 63.2 years is lower than all India average (Table 1) and its infant mortality is much higher at 54 per thousand live births compared to country’s average of 40 per thousand live births. Though its decadal growth rate in population at 17.1 percent is lower than all India average of 17.7 percent, yet its population density at 398 persons per square km., is higher than the country’s average. By contrast its urban population is less than half of all India average suggesting a domination of rural economy and lack of opportunities in urban areas for employment. A glance at the literacy figures suggest that its total literacy and male literacy respectively at 72.19 percent, 77.85 percent are lower than all India figures but the female literacy at 66.27 percent is higher than the corresponding country’s average of 64.64 percent. It is a poorer state with nearly one third of its total population (31.98 percent) and nearly 21 and 34 percent of its urban and rural population falling below poverty line (Table 1).

Table 1: Some demographic indicators: All-India and Assam state
Indicators	India	Assam
Expectation of Life at Birth, 2006-10(Male)	64.6	61.0
Expectation of Life at Birth, 2006-10(female)	67.7	63.2
Population (2011 census)	1210569573	31205576
Decadal Growth Rate (2001-2011)	17.7	17.1
Density(per sq. Km.)	382	398
Sex Ratio(per thousand male)	943	958
Urban population (%)	31.2	14.1
Literacy (person)	72.99	72.19
Male	80.89	77.85
Female	64.64	66.27
Birth rate	21.4	22.4
death rate	7	7.8
Infant mortality	40	54
Population below poverty (Rural)	25.70	33.89
Population below poverty (Urban)	13.70	20.49
Population below poverty (combined)	21.92	31.98

Model Specification

In the estimation of health system efficiency, our specification is based on a general stochastic frontier model that is presented as:

lnq_j = f(ln x) + v_j - u_j …………………(1)

Where: ln q_j is the health output (life expectancy or inverse of IMR) produced by a health system “j”

x is a vector of factor inputs represented by per capita health facilities (including per capita availability of hospital beds, per capita primary health centers (or sub centers or government medical institutions), per capita doctors, per capita paramedical staff, per capita skilled attention for birth.

v_j is the stochastic ( white noise ) error term.

u_j is a one-sided error term representing the technical inefficiency of the health system “j”

Both v_j and u_j are assumed to be independently and identically distributed (iid) with variance sv2 and su2, respectively

From the estimated relationship ln q^j = f (ln x) - u_j

The efficient level of health outcome (with zero technical inefficiency) is defined as:

ln q* = f (ln x)

This implies ln TEj = ln q^_j - ln q* = - u_j

Hence TE_j = e-u_j, 0<= e^-u_j<= 1

If u_j = 0 it implies e^-u_j = 1

Health system is technically efficient.

This implies that technical efficiency of j^th district health system is a relative measure of its output as a proportion of the corresponding frontier output.

A health system is technically efficient if its output level is on the frontier which in turn means that q/q* equals one in value.

Data Base

This study is based on secondary data. Information for the years 2012 and 2013 (the latest available year) is collected from various sources; including Health survey of India for the state[16], district portals of Assam and statistical abstract of the state.[15] At the district level main variables used from census and other government publications for this study are available for 23 districts due to creation of a number of districts at different points of time. Thus these 23 districts comprise our data set with main variables as infant mortality rates (IMR) and other parameters related to health infrastructure including number of primary health centers (PHCs), sub-centers (SCs), community health centers (CHCs), hospitals and dispensaries, total number of government health institutions, health manpower- medical and paramedical, and other variables relevant for depicting healthcare facilities, their utilization, health outcomes, socio-economic parameters like literacy levels in rural and urban areas, population growth, population density, urbanization and availability of basic amenities etc.

Statistical analysis tools used by our study include frontier regression technique applying STATA software.

Results and Discussion

The results of stochastic frontier model using panel data for 2012 and 2013 are presented in Table 2. All the variables are used in natural log. Results indicate a statistically significant influence for inverse IMR of two input variables namely total number of institutional deliveries carried out in all types of medical institutions and less than 24 hours stay after delivery in the health institution (Table 2). The negative sign of the latter variable depicts inadequate care after delivery leading to higher infant mortality whereas institutional delivery had a desirable positive impact on safe delivery and better chances of infant survival. Utilising variance of u and v (sigma u square and sigma v square given in last two rows) we find that resultant of u/v or inefficiency as higher than one, depicting the presence of inefficiency across districts. Table 3 presents actual and estimated IMR and district ranks in terms of achieving the estimated values of IMR. It could be observed that the district of Dhemji has the top rank in terms of achieving the estimated IMR. This is followed by Kamrup and Barpeta.. The lowest rank in terms of achieving the potential IMR is for Darrang followed by Kokrajhar (22nd rank) and Dhubri (21st rank). Generally high achievers have low actual IMR and low achievers have high actual IMR (Table 3). A further analysis for possible causes of better or lower performance is carried out by various factors presented in Tables 4-8.

Table 2: Stochastic Frontier results for Assam
Time-invariant inefficiency model Number of observations = 46 ; Number of groups = 23; Observations per group: min = 2; avg = 2; max = 2; Wald chi²(2) = 27.21; Log likelihood = 180.9165; Prob > chi² = 0.0000
Inverse log IMR	Coefficient	z values	P>\|z\|
Total institutional deliveries	0.0114	1.8800	0.0590
Less than 24hours stay after delivery in the health institution	-0.0064	-2.2400	0.0250
Constant	0.2523	7.2000	0.0000
mu	0.0281	4.6700	0.0000
lnsigma2	-9.1383	-28.8300	0.0000
ilgtgamma	3.7614	8.2100	0.0000
sigma2	0.0001
gamma	0.9773
sigma_u2	0.0001
sigma_v2	0.0000
Source: Estimated

Table 3: Actual and Estimated Infant Mortality rates (IMR) in Assam (district level)
Districts	Actual IMR(2012)	Estimated IMR	Achievement (estimated-actual)/estimated*100	Ranks based on achievements
Barpeta	43	37.77	-13.83	3
Bongaigaon	48	36.74	-30.64	4
Cachar	53	36.65	-44.62	10
Darrang	70	35.01	-99.97	23
Dhemaji	37	37.23	0.62	1
Dhubri	69	39.26	-75.73	21
Dibrugarh	51	33.16	-53.81	12
Goalpara	53	37.61	-40.92	7
Golaghat	56	35.48	-57.84	14
Hailakandi	52	36.83	-41.18	8
Jorhat	50	34.71	-44.04	9
Kamrup	39	34.77	-12.18	2
Karbi Anglong	60	38.25	-56.85	13
Karimganj	65	39.58	-64.23	16
Kokrajhar	74	37.98	-94.81	22
Lakhimpur	48	34.13	-40.63	6
Marigaon	63	36.66	-71.87	20
Nagaon	62	37.63	-64.75	17
Nalbari	58	33.78	-71.68	19
Dima Hasao*	54	35.15	-53.62	11
Sibsagar	56	34.27	-63.42	15
Sonitpur	61	35.78	-70.47	18
Tinsukia	50	36.27	-37.87	5
Source: Estimated; *earlier until February 2nd 1970 called as North Cachar Hills

A cursory glance at Table 4 indicates, for instance, that the top achiever (Dhemji) does not have the highest of either of total number of government medical institutions per lakh of population(23.16) or number of private institutions per lakh of population (.87) or total number of beds per lakh of population (41.39) (Table 4; row 6th). A further analysis of medical manpower both in rural and urban areas is presented in Table 5. It reveals some interesting pattern. For instance, the best achiever Dhemji has the highest number of urban doctors (140.83), rural and urban nurses (37 and 39.35 respectively) and rural and urban midwives (76.04 and 124.26 respectively). Comparing these figures of medical manpower availability with the low performing districts we find that Dhubri has indeed very low per lakh availability of the urban doctors (22.58), rural and urban nurses (14.72 and 7.36 respectively) and midwives (28.7 and 13.25 rural and urban respectively). Even Darrang, the lowest achiever also has nearly half of the per lakh availability of urban doctors (70.28), rural nurses (18.79) as well as rural and urban midwives (935.74 and 34.24). It is thus low availability of doctors, nurses and midwives in low performing districts that impedes the utilisation of medical institutions and beds.

Another important observation of a possible factor external to health care sector per se could be made from Table 6 which presents some basic population parameters both in rural and urban areas. It could be observed, for instance, that the decadal growth rate of population in 2001-11 has been higher in lowest performing district of Darrang and particularly its population density (586 per square km.) is more than double than the top achiever district of Dhemji. At the same time, urbanisation is lower in Darrang (5.98 percent). This pattern reveals another factor namely that with increasing rural density of population in low performing district of Darrang, the commensurate rise in medical manpower has not taken place and thus despite increasing need for care, utilisation of various facilities namely, medical institutions and beds, remained lower than optimum.

Table 4: Medical Institutions and number of Beds in Assam (Districts)
	Total no. of govt medical insts.	Total no. of govt medical insts. Per lakh population	No. Pvt med inst	No. Pvt med inst per lakh	Total Number of Beds	Total no. of beds per lakh population
Barpeta	326	19.25	10	0.59	582	34.36
Bongaigaon	127	17.19	21	2.84	458	61.99
Cachar	307	17.68	33	1.90	314	18.08
Darrang	215	23.16	13	1.40	392	42.22
Dhemaji	123	17.93	6	0.87	284	41.39
Dhubri	294	15.08	10	0.51	564	28.93
Dibrugarh	267	20.13	53	4.00	278	20.96
Goalpara	193	19.14	9	0.89	346	34.32
Golaghat	191	17.90	17	1.59	466	43.68
Hailakandi	121	18.35	0	0.00	220	33.37
Jorhat	196	17.94	38	3.48	628	57.50
Kamrup	317	20.89	153	10.08	680	44.81
Karbi Anglong	203	21.23	2	0.21	520	54.38
Karimganj	240	19.53	8	0.65	224	18.23
Kokrajhar	205	23.11	2	0.23	475	53.54
Lakhimpur	191	18.33	14	1.34	424	40.69
Morigaon	152	15.88	2	0.21	266	27.78
Nagaon	444	15.72	48	1.70	742	26.28
Nalbari	197	25.53	11	1.43	562	72.83
Dima Hasao	79	36.90	2	0.93	206	96.22
Sivasagar	269	23.37	25	2.17	494	42.92
Sonitpur	330	17.15	34	1.77	634	32.95
Tinsukia	194	14.61	33	2.49	342	25.75
Source: 15

Table 5: Medical Manpower (Rural and Urban; Assam Districts)
	Medical manpower Per lakh population
	NOF Doctors Rural	NOF Doctors Urban	NOF Nurses Rural	NOF Nurses Urban	NOF Midwives Rural	NOF Midwives Urban
Barpeta	8.41	14.25	17.40	16.97	35.18	33.25
Bongaigaon	17.49	17.30	24.01	25.50	48.49	59.19
Cachar	12.38	14.58	7.25	4.75	14.42	6.97
Darrang	12.49	70.28	18.79	21.62	35.74	34.24
Dhemaji	26.65	140.83	37.00	39.35	76.04	124.26
Dhubri	11.34	22.58	14.72	7.36	28.70	13.25
Dibrugarh	13.30	18.87	5.63	17.64	27.90	6.56
Goalpara	12.64	24.63	11.95	10.86	12.87	12.31
Golaghat	16.10	44.00	14.65	27.63	28.99	50.14
Hailakandi	24.54	56.09	26.02	51.93	49.74	49.85
Jorhat	4.82	6.35	11.70	9.98	23.98	8.16
Kamrup	21.82	36.52	26.32	36.52	47.41	82.17
Karbi Anglong	35.57	40.72	31.90	32.75	69.25	15.93
Karimganj	9.29	32.82	18.32	22.79	39.41	38.29
Kokrajhar	18.02	58.24	18.26	36.40	37.49	49.14
Lakhimpur	21.03	35.04	25.77	20.80	51.75	27.37
Morigaon	8.60	30.01	9.05	30.01	16.97	36.84
Nagaon	4.12	10.01	0.45	4.60	12.71	5.14
Nalbari	22.64	22.97	31.79	26.59	60.24	42.31
Dima Hasao	80.47	75.21	31.00	16.00	65.96	24.00
Sivasagar	16.24	22.71	13.83	22.71	25.17	45.41
Sonitpur	5.26	8.63	5.83	18.41	10.86	21.28
Tinsukia	4.89	14.35	6.21	4.91	12.70	8.31
Source: 15; NOF=number of

A further factor external to health care sector could be discerned from Table 7 which presents literacy levels. It is notable, for instance, that most of the higher performing districts like Dhemaji, Kamrup and Barpeta, the literacy levels particularly total literacy, male and female literacy and urban literacy are higher than the low performing districts of Darrang, Kokrajhar and Dhubri (Table 7). Thus literacy levels also seem to have an impact on utilisation of medical facilities. Higher literacy is associated in general with higher levels of utilisation.

Another factor to explain variation in efficiency is noted in terms of various basic facilities of electricity, water and latrines (Table 8). An important observation emerging from Table 8 in this regard is that larger road length also seems to be a positive factor for Dhemji in relation to low performers like Darrang and Dhubri, whereas low availability of households latrines seemed to be associated as a negative factor for low performer namely, Kokrajhar.

Thus in our second stage of estimation we tried all these variables to explain the residual efficiency, none of these variables, however, appeared statistically significant and thus the results are not presented on this stage of estimation in the paper.

Table 6: Population Assam districts (2011 census): Growth, rural and urban shares
District	% share to total Popula- tion, 2011	Decadal Growth Rate 2001-11	Population density2011	Rural percent of Population	Urabanisation
Barpeta	5.43	21.43	742	91.3	8.7
Bongaigaon	2.37	20.59	676	85.14	14.86
Cachar	5.56	20.19	459	81.83	18.17
Darrang	2.98	22.19	586	94.02	5.98
Dhemaji	2.2	19.97	212	92.96	7.04
Dhubri	6.25	24.44	896	89.55	10.45
Dibrugarh	4.25	11.92	392	81.62	18.38
Goalpara	3.23	22.64	553	86.31	13.69
Golaghat	3.42	12.75	305	90.84	9.16
Hailakandi	2.11	21.45	497	92.7	7.3
Jorhat	3.5	9.31	383	79.81	20.19
Kamrup	4.86	15.69	489	90.62	9.38
Karbi Anglong	3.06	17.58	92	88.19	11.81
Karimganj	3.94	21.9	679	91.07	8.93
Kokrajhar	2.84	5.21	269	93.81	6.19
Lakhimpur	3.34	17.22	458	91.24	8.76
Morigaon	3.07	23.34	617	92.34	7.66
Nagaon	9.05	22	711	86.91	13.09
Nalbari	2.47	11.99	733	89.28	10.72
Dima Hasao	0.69	13.84	44	70.81	29.19
Sivasagar	3.69	9.44	431	90.44	9.56
Sonitpur	6.17	15.55	370	90.96	9.04
Tinsukia	4.26	15.47	350	80.06	19.94
Assam	100	17.07	398	85.9	14.1
Source: 15

Table 7: Literacy in Assam( 2011 census); Total, male, female, rural and urban(district level)
District	D.G.R. 2001-11	Total Literacy	Male literacy	Female literacy	Rural literacy	Urban literacy	male-female literacy gap (MFLG) Total	MFLG Rural	MFLG Urban
Barpeta	16.15	63.81	69.29	58.06	61.47	86.28	11.23	11.53	8.5
Bongaigaon	12.96	69.74	74.87	64.43	66.42	87.37	10.44	10.83	8.25
Cachar	56.65	79.34	84.78	73.68	77.08	87.39	11.1	12.37	5.96
Darrang	33.07	63.08	67.87	58.04	61.5	85.92	9.83	9.89	8.33
Dhemaji	24.34	72.7	79.84	65.21	71.81	84.02	14.63	15.04	9.11
Dhubri	5.85	58.34	63.1	53.33	55.25	82.28	9.77	9.81	9.77
Dibrugarh	6.69	76.05	82.82	68.99	72.75	88	13.83	15.55	5.95
Goalpara	106.36	67.37	71.46	63.13	65.93	76.08	8.33	8.33	8.44
Golaghat	20.46	77.43	83.56	71.09	75.94	91.74	12.47	13.22	5.14
Hailakandi	9.2	74.33	80.74	67.6	72.73	92.93	13.14	13.94	4.77
Jorhat	28.73	82.15	87.63	76.45	80.01	72.5	11.18	12.66	5.17
Kamrup	145.16	75.55	81.3	69.47	74.21	87.89	11.83	12.14	9.36
Karbi Anglong	22.88	69.25	76.14	62	66.69	87.37	14.14	14.9	8.64
Karimganj	48.54	78.22	84.12	72.09	76.66	92.82	12.03	12.84	4.95
Kokrajhar	6.68	65.22	71.89	58.27	63.63	87.86	13.62	13.92	8.53
Lakhimpur	40.18	77.2	83.52	70.67	76.22	86.93	12.85	13.36	7.47
Morigaon	92.95	68.03	71.9	64.04	66.6	84.17	7.86	7.83	7.98
Nagaon	32.79	72.37	76.51	68.07	69.96	86.34	8.44	8.7	6.82
Nalbari	200.99	78.63	84.36	72.57	77.22	89.89	11.79	12.29	7.84
Dima Hasao	5.14	77.54	83.29	71.33	71.13	92.24	11.96	13.84	6.67
Sivasagar	13.29	80.41	85.84	74.71	79.27	90.92	11.13	11.82	4.1
Sonitpur	-1.11	67.34	73.65	60.73	64.98	81.65	12.92	13.66	5.93
Tinsukia	18.21	69.66	77.19	61.73	65.05	82.08	15.46	17.2	7.72
Assam	27.89	72.19	77.85	66.27	69.34	88.47	11.58	12.37	6.87
Source: 15

Table 8: Availability of water, latrines, electricity and roads in Assam (Districts)
District	percent household having (PHHH) water within premises	PHHH water near premises	PHHH electricity as source of lighting	PHHH latrine facility within premises	Number of Villages (NOV) electrified (2012-13)	NOV electrified (Cumulative)(2013-14)	Road length (RL)Highway	RL district	RL rural	RL urban	RL total
Barpeta	65	22.1	25.1	71.9	998	998	158	171	1463	31	1824
Bongaigaon	69.1	18.8	33.2	56.4	838	844	41	16	712	36	805
Cachar	21.8	46.2	38.1	81	890	890	107	165	848	42	1163
Darrang	58.7	23.1	24.2	50.6	1305	1305	143	119	785	20	1066
Dhemaji	54.9	29.4	21.8	44.8	1027	1027	45	42	1168	86	1340
Dhubri	65.2	19.6	17.4	43.3	1226	1226	56	53	1007	38	1154
Dibrugarh	67.6	25.6	50.1	78.7	1035	1035	155	163	1357	63	1738
Goalpara	58.6	22.3	39.8	65.5	741	741	137	48	1354	10	1550
Golaghat	46.1	28.6	36.6	68.4	1032	1031	160	157	2131	33	2480
Hailakandi	10.7	52.3	30.7	83.1	313	306	17	99	351	6	473
Jorhat	44.9	35.1	52.4	66.7	769	769	162	89	1714	86	2051
Kamrup	64.8	20.2	40	59.2	1293	1304	89	230	2916	0	3235
Karbi Anglong	35.2	33.4	30.5	54.8	2252	2256	337	561	3341	70	4309
Karimganj	17.9	48.8	28.7	85	758	759	35	342	539	19	936
Kokrajhar	57.5	23.9	23	29.2	806	829	61	131	1632	16	1841
Lakhimpur	56.9	25.6	29.7	58	1087	1093	121	97	828	44	1091
Morigaon	69.4	18.4	28.2	60.1	474	484	142	86	909	13	1150
Nagaon	64.6	19.9	34.9	75.3	1327	1327	297	326	2409	88	3120
Nalbari	70.5	18.7	44	66.7	798	798	119	45	779	22	965
Dima Hasao	17.8	35.7	45.2	69.4	482	496	380	199	1237	60	1876
Sivasagar	49.7	35.2	50	74.2	452	452	100	310	2425	40	2874
Sonitpur	53.7	27	34.5	59.2	1542	1542	55	402	2133	59	2648
Tinsukia	66.2	25.9	60.3	81.1	1075	1075	63	171	1479	52	1765
Assam	54.8	26.7	37	64.9	22520	22587	3134	4413	36544	1409	45500
Source: 15

Conclusions

Utilizing a stochastic frontier model our results thus indicate that better availability of medical manpower including doctors, nurses and midwives has led to an optimum utilization of existing medical institutions and beds in some of the districts in Assam. Thus based on achievement of their existing normative potential, districts like Dhemji, Kamrup and Barpeta are three top ranking districts. The low performing districts based on these criteria, including Darrang, Dhubri and Kokrajhar have been not able to utilize existing medical institutions and beds capacity due to the constraint of inadequate medical manpower, higher population density, higher rural populations, lower literacy levels and lack of comparable roads development relative to efficient districts in the state. Thus the overall efficiency in the district level health system could be improved to bring down infant mortality in lower performing district with more focus of policy on the factors that have been highlighted by our study.

References

W.H.O. The World Health Report, 2000: Health Systems: Improving Performance. World Health Organisation. 2000.
Murray CJL, Frenk J. A WHO Framework for Health System Performance Assessment: Global Programme on Evidence and Information for Policy. World Bank.1999.
Worthington AC. Frontier Efficiency Measurement in Health Care: A Review of Empirical Techniques and Selected Applications. Medical Care Research and Review. 2004;61(2):135-170.
Hollingsworth B, Wildman J. The Efficiency of Health Production: Re-estimating the WHO Panel Data Using Parametric and Nonparametric Approaches to Provide Additional Information. Working Paper No. 131, Centre for Health Programme Evaluation, Monash University, Australia. 2002.
Jamison DT, Sandbu M, Wang J. Cross Country Variation in Mortality Decline, 1962-87: The Role of Country Specific Technical Progress. CMH Working Paper Series Paper No WGI: 4. Commission on Macroeconomics and Health, WHO. 2001.
Salomon JA, Mathers CD, Murray CJL, Ferguson B. Method for Life Expectancy and Healthy Life Expectancy Uncertainty Analysis: Global Programme on Evidence for Health Policy. Working Paper No 10. World Health Organization. 2001.
Evans DB, Tandon A, Murray CJL, Lauer JA. The Comparative Efficiency of National Health Systems in Producing Health: An Analysis of 191 Countries. GPE Discussion Paper Series: No. 29. EIP/GPE/EQC/World Health Organization. 2001.
Wang J, Jamison DT, Bos E, Preker A, Peaboy J. Measuring Country Performance on Health: Selected Indicators for 115 Countries. Human Development Network. Health, Nutrition and Population Studies. IBRD/World Bank, Washington. 1999.
Shankar D, Kathuria V. Health System Performance in Rural India. Efficiency Estimates across States. Economic and Political Weekly. 2004, March 27; 1427-1433.
Fried HO, Lovell CA, Schmidt SS. The Measurement Of Productive Efficiency: Techniques And Applications. Oxford University Press, New York.1993.
Charnes A,Cooper WW, Lewin AY, Seiford LM. Data envelopment analysis:Theory, Methodology And Applications. Kluwer, Boston.1995.
Coelli T, Rao DSP, Battese G. An Introduction To Efficiency And Productivity Analysis. Kluwer, Boston. 1998.
Hollingsworth B, Dawson PJ, Maniadakis N. Efficiency Measurement Of Health Care: A Review of Non-Parametric Methods and Applications. Health Care Management Science.1999;2(3):161-72.
Purohit BC. Efficiency of Social Sector Expenditure in India. Routledge/ Francis and Taylor, UK. 2014.
Government of Assam. Statistical Hand Book 2014. Directorate of Economics and Statistics, March, Guwahati. 2015.
Government of India. Annual Health Survey Bulletin, 2012-13, Office of the Registrar General 8t Census Commissioner, Ministry of Home Affairs, New Delhi. 2015.