Cost of stroke from a tertiary center in northwest India
Correspondence Address: Source of Support: Department of Neurology intramural research fund, Conflict of Interest: None DOI: 10.4103/0028-3886.125270
Source of Support: Department of Neurology intramural research fund, Conflict of Interest: None
Aim: We aimed to study the cost of stroke, its predictors, and the impact on social determinants of the family. Settings and Design: This prospective study was done in the Stroke unit and Neurology clinic between April 2009 and October 2011. Materials and Methods: All first ever stroke patients during the study period were enrolled. Direct and indirect costs at admission, at 1 and 6 months follow-up were obtained. The follow-up included information about the patient's poststroke outcome using modified Rankin Scale (mRS), work status, modifications made at home, loan requirement, etc., Results: Two hundred patients were enrolled in this study and final analysis was performed on 189 patients. The mean age was 58 ± 13 years and 128 (67.7%) were men. Majority (54%) were living in a joint family. The mean overall cost of stroke per patient was rupees (INR) 80612 at 6 months. Higher income (P = 0.008), poor outcome (mRS >2) (P = 0.001), and length of hospital stay (P = 0.001) were the cost driving factors of total cost of stroke at 6 months. There was a decline in the requirement of help (P < 0.0001) and need for loan (P = 0.003) at 6 months follow-up. Conclusions: Direct medical cost or acute care of stroke accounted for a major component of cost of stroke. Poor outcome, length of hospital stay, and higher income were the cost driving factors. The socioeconomic impact on the family decreased at follow up probably due to joint family system.
Keywords: Cost, developing country, indirect costs, socioeconomic, stroke, treatment
Stroke is one of the leading causes of death and disabilities worldwide. There is a high economic impact of stroke in developed countries like Australia, United Kingdom, Canada, New Zealand, Korea, and Taiwan. ,,,,, In India, a good number of stroke epidemiological data have been published recently. ,, However, little is known about the cost and economic impact of stroke. The health system in India is also different from other countries. While Government hospitals are available, the major health provider is the private sector. Health insurance covers only 10% of people and old people are not covered by health insurance policies.  The cost of treating a stroke patient can thus differ from other regions of the world. We carried out this study with the following objectives: a) to study the cost of illness of stroke, that is, direct medical, nonmedical and indirect costs; b) To study the predictors of cost of stroke; and c) to explore the changes in the social determinants following stroke in the family.
This prospective study was carried out in the Stroke unit and Neurology clinic of a tertiary referral center in northwest India from April 2009 to October 2011. The study protocol was approved by the Institutional Research Committee.
Stroke was defined as per World Health Organization (WHO) criteria and was confirmed by computed tomography (CT) or magnetic resonance imaging (MRI). Both ischemic as well as hemorrhagic stroke cases were included in the study. An informed consent was taken from all patients before the questionnaire was administered. If the patient had cognitive impairment, altered sensorium or dysphasia, then the consent was obtained from the next of kin.
The inclusion criteria were: a) all consecutive first ever stroke patients; b) age ≥ 18 years; ischemic and hemorrhagic stroke patients; and c) patients with premorbid modified Rankin Scale (mRS) of 0-1. Exclusion criteria were: a) patients with subarachnoid hemorrhage; and b) patients with one or more chronic diseases like renal failure, liver disease, bilateral severe osteoarthritis of knee, hip, etc.
Baseline information included demographics, medical history, Oxfordshire community stroke project classification (OCSP), prestroke mRS, and risk factor assessment. The questionnaire also included information about the number of earning members, family income, and the patient's housing conditions prior to the stroke. The cost-related data collected was analyzed from a societal perspective and included the direct as well as indirect costs. Intangible costs, which refer to patients' psychological pain and discomfort, were not included. The information related to hospital cost was obtained from the hospital bills. Patients and their caregivers provided the information regarding family income and expenditures related to hospitalization both during admission and follow-up. The same research staff had collected the information throughout the study.
Direct cost was calculated as a sum of direct medical and nonmedical cost.
Direct medical costs
The information related to cost of hospitalization, laboratory, radiology and cardiology-related investigations, drugs, nursing charges, consultant fees, rehabilitation services, and interdepartmental consultations was obtained from the hospital bills.
Direct nonmedical costs
Nonmedical direct costs include transportation costs to healthcare providers; relocation expenses; and costs of making changes to one's diet, house, car, or related items.  This study included the cost of transportation of the patient to the hospital and costs incurred by the caregivers during the period of patient's hospitalization: meals, transportation, lodging, etc., This was based on the information provided by the patient and/or the caregivers. If the patient was brought to the hospital in an ambulance or hired vehicle, the actual charges were used. If the patient was transported using own or friends'/relatives' vehicle, then the cost was calculated using the approximate distance and average fuel cost. If the patient had to be shifted to another house or additional appliances were purchased, then those costs were added to direct nonmedical costs.
Indirect costs of disease are defined as the production value lost to society due to absence from work, disability and death.  In this study, the human capital approach was used to estimate the indirect cost or productivity losses.  The indirect costs were estimated for patients who had a paid job prior to the stroke and also for patients involved in unpaid domestic work. In case of sick leave, the number of sick leave days was multiplied by the average daily income of the patient. Productivity losses were also calculated for the informal caregivers based on the number of working days missed. For the patients or caregivers who performed unpaid domestic activities, the loss of productivity was based on the average labor wages applicable in our country.  For patients who had a premature death, productivity losses were estimated from the time of their death up to 6 months of follow-up.
The patients were followed up at 1 and 6 months' time following the stroke. At the time of follow up, direct medical costs related to hospital record charges, laboratory investigations, drugs, and rehabilitation were calculated. The direct nonmedical costs were based on the transportation charges for the patient's hospital visit and relocation expenses, if any. For the indirect costs, it was assumed that there was loss of one day of paid production for the patient as well as for the caregivers and calculations were made accordingly. A telephonic follow-up was conducted if any patient did not come to the hospital on the designated day. In addition, information regarding patient's functional and socioeconomic status following the stroke was obtained. Patient's functional capacity was assessed by getting information about the patient's poststroke mRS, work status, need for a caregiver either as a family member or a paid helper, modifications made at home or appliances purchased for the patient following the stroke. If the patient needed help for daily activities, information about the number of caregivers and the impact of the patient's stroke on their earning capacity was also obtained.
To assess the economic impact of stroke on the patient and his/her family, there were questions about any change in the patient's housing condition, and also if the patient's family had to borrow money for the treatment in the form of a bank loan or a soft loan from friends/relatives.
Joint family (extended family/complex family) was defined as multiple generations in a family living in a single house. Education was classified into lower (illiterate and up to primary level) and upper (secondary school and beyond). Income was classified into lower ( < Rs 15,000 per month) and upper ( ≥ Rs 15,000 per month).
The statistical analysis was performed using SPSS version 21 (Armonk, NY: IBM Corp.). As the cost variables were skewed, nonparametric bootstrap algorithm was used to calculate 95% confidence intervals (CI). The method used to calculate 95% CI was a bias-corrected and accelerated bootstrapping method with B = 1000 bootstrap replications. Comparisons of mean costs between groups were tested using a bootstrap t and ANOVA test. Chi-square and Fisher's exact tests were used to study the relationship between categorical variables. Multiple linear regression was performed to find out the cost driving factors. The variables included in the model were income, OCSP, outcome, length of hospital stay and complications. A P < 0.05 was considered significant.
Demographic and social parameters
Two hundred patients were enrolled in this study. We included 189 patients in the final analysis (excluded 11 patients because of incomplete information at the time of follow up; not possible to contact). The mean age was 58 ± 13 years and 128 (67.7%) were men. Out of 189 patients, 102 (54%) were living in a joint family and majority of the patients 137 (72.5%) had two or less than two earning members in their family and 114 (60.6%) belonged to lower income category (≤Rs 15,000 per month). The mean length of hospital stay was 13 ± 8 days. Detailed description of demographic and social parameters are shown in [Table 1].
Cost of stroke at the time of admission and follow up
The mean overall cost of stroke per patient in our study was Rs (INR) 80612 (95% CI: 72574-88574) at 6 months. The mean overall direct medical, and nonmedical and indirect cost was Rs 52555 (95% CI: 47319-58312), Rs 4826 (95% CI: 4122-5571) and Rs 23230 (95% CI: 20007-26775), respectively, at 6 months. Sixty-five percent of the total cost was due to direct medical cost, 6% due to direct nonmedical cost and indirect cost contributed 29% to the total cost of stroke in our study. The detailed description of total cost at the time of admission and follow up is shown in [Table 2].
Comparison of total cost at 6 months with demographic, social, and clinical parameters using univariate and multiple regression analysis
In the univariate analysis, the following factors were associated with higher cost: higher income (P = 0.007), OCSP (total anterior circulation stroke (TACS), partial anterior circulation stroke (PACS)) (P = 0.01), poor outcome (P = 0.001), length of hospital stay (P = 0.001) and complications (pneumonia, shoulder subluxation, urinary tract infection (UTI), seizure, coronary event, fractures) (P = 0.04) [Table 3]. There was a significant difference seen in the cost of care according to outcome, patients who had a severe stroke had higher cost of care [Table 3]. There was no significant difference was seen in total cost according to age (P = 0.85), gender (P = 0.46), marital status (P = 0.59), religion (P = 0.62), family type (P = 0.99), number of earning members (0.63), type of stroke (0.69), and health insurance (P = 0.21).
Only higher income (P = 0.008), poor outcome (P = 0.006), and length of hospital stay (P = 0.001) were the cost driving factors in the multiple regression analysis [Table 4].
Changes in the social determinants during follow up
At 1 month follow up data from 161 patients (died: 38; not possible to contact: 1) and at 6 months follow up, data from 134 patients (died: 17; not possible to contact: 10) were available. The stroke recovery was better at 6 months as compared with 1 month (P < 0.0001). The need of help by stroke patients (P < 0.0001), the impact on helper's income (P = 0.01), and the need for loan (P = 0.003) were reduced during follow-up [Table 5]. A small proportion of patients (1 month 12.6% vs. 6 months 9%; P = 0.33) used alternative medicine, which increased the treatment cost at 6 months (1 month Rs 1397 ± 999 vs. 6 months 2844 ± 2512; P = 0.05).
We studied the cost (direct medical, nonmedical, and indirect) and socioeconomic impact of stroke in northwest India. In the multiple regression analysis, poor outcome, length of hospital stay, and higher income were the cost driving factors at 6 months follow up. During follow up at 6 months' time, the stroke recovery improved and the changes in the patient's social life like need for help, impact on helper's income and need for loan decreased as compared with 1 month.
In our study, the overall cost of stroke (ischemic and hemorrhagic stroke) was Rs (INR) 80612 (US $1520). Patients stayed for a longer time in the hospital in our study as compared with studies from Taiwan  and Singapore  but shorter than a study conducted in China.  The length of hospital stay in our study was longer because of our intensive ongoing in-hospital rehabilitation program. 
The main component of cost of stroke in our study was direct medical cost (65%). Only about 10% of the population in India has health insurance,  whereas in a Chinese study, 62% of patients had health insurance.  Hence, the health costs in a vast majority of our patients are out of pocket expenses. Higher income, poor outcome, and length of hospital stay were the main factors. The indirect cost of patients who belonged to higher income group was high because their wages were high and also they availed the private ward facilities for their treatment so their total cost was high. Patients with poor outcome had higher cost because they stayed for longer period of time in the hospital, due to more medical complications and need for extra care.
The present cost of stroke can be reduced by decreasing the length of hospital stay. This could be achieved by implementing early supported discharge and home-based rehabilitation.  The proposed ATTEND trial (CTRI/2013/04/003557, ACTRN12613000078752) comparing caregiver led home-based rehabilitation versus conventional rehabilitation in India will answer this important question.
At 6 months, changes in work situations of the patients were less as compared with 1 month. Also, at 6 months, because of the improvement, very few patients required help. Furthermore, the relatives were the main helpers, and so the impact on their income reduced with time. In addition, the living situation did not change as majority of the patients returned back to their homes unlike other developed countries. In our study, 27.5% of patients had more than two earning members in their family and half of the cohort was living in a joint family. This was probably the reason why there was not much impact in the living and social situation of the patients.
The proportion of patients who were taking loan for their treatment was very low (11%). The reasons for this were: the per capita income of Punjab State is Rs 68,998 and Ludhiana city ranks first in the Punjab Human Development Index (2001).  Only 5.2% population of Punjab lives below poverty line.  However, a study from China showed that one-third of patients and their families had fallen below the poverty line soon after stroke. 
A limitation of this study is it is a hospital-based report from a single center in northwest India. Our data may represent the cost of stroke in a private sector hospital in India. As this is not a population-based study, the data has not been compared with the data from other countries. Despite this limitation, this study, for the first time has explored the cost and socioeconomic impact of stroke from a developing country in a comprehensive way. Multi centered studies across the country are warranted to find out the cost of stroke in India.
To conclude, the direct medical cost or acute care of stroke in the hospital accounted for a major component of cost of stroke in our study. The cost driving factors were higher income, poor outcome and length of hospital stay. There was a decrease in socioeconomic impact at follow up. With the help of early supported discharge and home-based rehabilitation, we may be able to reduce the length of hospital stay and hence the acute care cost.
The authors sincerely thank Mrs. Madhu Bala for the data entry.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]