A hospital based cross-sectional study was conducted in an Outpatient Department of Urban Health Training Centre, Srinagar, Uttarakhand and Rural Health Training Centre, Kritinagar, Uttarakhand, India over a period of four months (from September to December 2015). These health centers belong to the Department of Community Medicine of Veer Chandra Singh Garhwali Government Medical Sciences and Research Institute, Srinagar, Uttarakhand.

All the patient’s attending the outpatient department constituted the eligible study population. The sample size was determined following the WHO guidelines for medicine [30]. With alpha error of 20%, the sample size came to be 150. Taking non-response rate of 20%, the final minimum sample size was 180. A total of 200 patients were studied as a subcomponent of a primary study conducted on patient satisfaction in these health centres. An interview schedule was prepared and pretested. A prior approval from the Institutional Ethical Committee was obtained and informed written consent from patients was taken. A random sampling was done to enroll the required number of patients in the study. They were examined, about the diseases they presented with in the health facility. A semi-structured pretested questionnaire was used to interview the patients. The questionnaire was pretested with 10 patients before starting the study. The questionnaire was finalized after incorporating grammar changes based on the pretesting. There were 17 open-ended questions to study demographic details of the patients, diseases they presented with in the health facility, OOP expenditure like travel cost, drug cost, investigation cost and others. This was translated in the local language i.e., Hindi for better understanding of the study participants. Hindi version of form was back translated in English to check the validity of the form. A trained investigator interviewed patients according to the questionnaire. Each patient interview took approximately 15 minute to half hour.

OOP expenditure was calculated as the total amount of money spent by the respondents on both medical expenses (registration fees, medicines, and investigations) and non-medical expenses (transport, and others) for an episode of illness.

Data analysis was done using SPSS 16.0, windows compatible software. The large number of participants reported zero OOP expenditure in the present study. In addition, our positive OOP expenditure results were right-skewed. To overcome these effects, we used the Wilcoxon rank-sum test to study differences in positive OOP expenditure in binary categorical variables (e.g., sex, age, education, occupation family income etc.,) and Kruskal Wallis in more than two category variables (e.g., presenting diagnosis). The hierarchical stepwise multiple linear regression analysis used the total OOP expenditure as the dependent variable. Because the total OOP expenditures was not normally distributed, the dependent variable was log-transformed for the analysis. A probability p value of < 0.05 was taken for statistical significance.

A total of 210 patients were invited to participate in the study, out of which 200 participated in the study; therefore the overall response rate was 95%. Age wise, the majority (54.0%) belonged to more than 35 yrs. The mean age of the studied patients 40.36 (SD 19.45) in years. Sex wise, the study subjects were almost equal {43.5% (male) vs. 56.5% (female)}. Average annual family income was 11718.5±13457.3 INR of the total study subjects, 42.5% were unskilled or had no occupation (unemployed, student, and homemaker). The majority had travelled less than or equal to 10 km to seek health care. Half of the study subjects presented with communicable diseases [Table/Fig-1].

The health expenditure of the outpatient revealed that major part of the expenditure is the direct cost of the treatment, including registration fee, drugs, and investigation. Mean OOP expenditure for drugs and investigations was 509 INR and 673.1 INR, respectively. Mean total OOP expenditure was 303.1 INR. The proportion of monthly family income spent on OOP expenditure was 2.58% [Table/Fig-2].

Total monthly family income, occupation, and presenting diagnosis were observed to be significantly associated with health expenditure on medicine/drugs. Statistically significant difference (p<0.05) was observed in overall OOP expenditure and expenditure on investigation, for the type of health facility and presenting diagnosis [Table/Fig-3].

In the hierarchical stepwise multiple linear regression analysis, demographic factors (Block 1, [Table/Fig-4]) explained 1.7% of the variance (adjusted R²=0.017) in total OOP expenditure when none of the other factors were controlled for. Diagnosis and socioeconomic variables (Block 2, 3, [Table/Fig-4]) explained an additional nearly 3.6%, and health facility set up an additional 6% of the variance (Block 4, [Table/Fig-4]). Altogether, the final model explained 9.7% of the variance in total OOP expenditure [Table/Fig-4].

The present facility based cross-sectional study revealed the out of the expenditure pattern in seeking primary health care in the hilly area of Garhwal, Uttarakhand. Nearly 90% of the studied patient reported out of expenditure in direct or indirect form. OOP expenditure of the outpatients in primary health care was observed to be 2.58% of their monthly family income. Comparatively, a census conducted in Bengaluru, India reported overall, 69.6% of households made OOP payments, median of 3.2% of their total income for outpatient care [13]. In addition, a longitudinal community based study conducted in rural Maharashtra, India 2004-2005, observed OOP expenditure to be 4.3% of their annual income [31]. On the contrary community based, National sample survey 2004, 32.9% of the households reported OOP payments for outpatient care, and the mean and median amounts paid OOP were US$ 9.4 and US$ 4.2, respectively [23]. Another study conducted in under-five children of urban slums Puducherry, India found the proportion of family income spent toward OOP was 6.7%, which may be due to the different study settings [32].

Mean and Median OOP in this study was 303.1 INR (4.5 US $) and 60 INR (0.9 US$), respectively. Against expected, free primary care, services would reduce OOP payments in a hospital outpatient department. A cross-sectional study conducted between 2009-2010 covering all of the households in urban Bengaluru found that, the monthly median OOP payment of outpatient care was 400 INR (95% CI=380-403.5) (USD 8.1) [13]. The prime reason for this is different methodology adopted by the two studies. The study conducted in urban Bengaluru was a community based study and it measured OOP expenditure by individuals/households within 30 days [13].

In the present study, it was found with age more than 35 years the OOP on drugs (575.8 INR, 8.1 US$) increases, similar to the findings a prospective survey conducted to assess costs of Acute Respiratory Infection (ARI) episodes in north India also reported that OOP inpatient cost was the highest among adults aged more than or equal to 65 years [20]. Higher OOP expenditure among elderly could be due to more episodes of illnesses or multiple chronic conditions. It was also noticed that patient with family income more than 10000 INR per month and unskilled occupation are significantly more likely to spend on drugs.

The mean OOP expenditure on drugs and investigations was found to be 509 INR (7.6 US$), and 673.1 INR (10.0 US$) respectively in the present study. In 2012, medicines constituted the largest proportion of OOP expenditure in many other studies [16,33]. The patient consulted in Urban Health Training Centre had a significantly more chance to pay out for drugs (mean 546.6 INR, 7.6 US$), private investigation (955.2 INR, 13.4 US$) and thereby have an overall higher OOP (602.4 INR, 8.4 US$). The key reason of not using government supplies of medicine could be a patient’s perception of quality. In addition, there is easy availability of chemist and private laboratory in urban areas, which a rural patient lacks. This may be also because of lack of supplies and medicine and reagent for investigation in government health facility. An evaluation in Punjab and Haryana stated that the drug availability in public health care sector varied from 45%-51% [19]. A study of 2011 conducted in rural districts of Tajikistan was found, the median total expenditure of the users of primary healthcare facilities was 8.8 US$ and the median amount spent on drugs was 10.7 US$ [18]. This may be because in our study the medical college provided free OPD services through its Urban and Rural Health Training Centre facility.

The present study also documented that the patients presented with non-communicable diseases were significantly more OOP on drugs (771.1 INR, 10.8 US$) and investigation (1069.7 INR, 15.0 US$), leading to higher total OOP (783.0 INR, 11.0 US$). This is because non-communicable diseases are usually chronic and requires a life long treatment. In the present study, non-communicable disease patients spent 6.5% of their monthly family income on health care. A study found 4.8% OOP spending on drugs and pharmaceutical products; accounted due for cardiovascular-related chronic conditions [25]. The median OOP payment per chronic condition was 320 INR (6.5 US$), with the greatest share on direct medical care in a cross-sectional study in urban Bengaluru [13]. Another study conducted in urban households in West Bengal, found less OOP expenditure, 4.2% of their annual household expenditure on care for chronic conditions [34]. This difference is possibly explained by inflation in INR over the years.

The present study found that mean OOP payment for communicable disease was 1.6 US$. A study conducted during 2012-2013 at 24 health care facilities covering all three levels of health care in the National Capital Region and Srinagar, Jammu & Kashmir, India, reported the total direct cost of ARI episodes requiring outpatient expenditure care was 4US$-6$ for public institutions [20]. This difference accounts for the difference in the study setting. Srinagar, Uttarakhand is a low tier city compared to National Capital Region and Srinagar, Jammu & Kashmir, India.

The primary limitation of the present study is that we had a single visit recall period, which could lead to underestimation of OOP expenditure.

The implementation of primary health care needs to be further strengthened in order to improve access of public sector hospitals for curative care. There is a current need for, improvements in the availability of medications and diagnostics within the government health centres and the control of the costs of such services in the private sector are needed.