We estimated the cost per DALY averted of a surgical unit of a 106-bed for-profit private hospital in north India. The hospital was selected based on cooperation, data availability and accessibility. It is unlikely that the government health facilities in India will keep computerized patient records, only the private facilities generally do that. As a detailed study like this requires patient-level data, it is not possible to review all case records manually; hence, we chose the private hospital that provided support in extracting their computerized patient records without identifiers. In India government health care services are provided in several levels starting with sub-centres, primary health care centres and community health centres to provide primary care. The district hospitals of 150-bed provide secondary care and there are tertiary care hospitals. Along with these, a substantial proportion of care is provided by the private providers and there are non-profit charitable hospitals as well. The private sector delivers about 80 % of outpatient care and 60 % of all inpatient care in India [19]. The private hospitals can range from few bedded nursing homes to multi-speciality hospitals. The 106-bed for-profit study hospital is like a small district hospital in government setting providing secondary care. We admit that it is not a representative hospital in a country like India, however, the cost per DALY averted for surgical unit in this hospital compare favourably with the same for several other interventions documented in the literature [20, 21]. It should be noted in this context that majority of the surgeries performed by the study hospital were related to maternal conditions and the cost per DALY averted is driven by this case-mix, however, this is common in small hospitals in India. In another Indian study on cost of surgical procedures also noted that the majority of the surgeries performed in a 200-bed government district hospital were lower section caesarean sections followed by hysterectomy, appendectomy and hernia repair [22]. Even though our results are highly context specific, our study will encourage more research to prove that surgical interventions are cost-effective.
The costs per visit (USD 17) and bed-day (USD 54) at the study hospital were lower than those reported for a private hospital in a recent hospital costing study in India (USD 42 and USD 134, respectively, at 2011 prices) [14]. The inpatient department of the study hospital consumed more resources than the outpatient department, as is typical for hospitals in India. The unit costs at the operating theatre (USD 417) and labour room (USD 403) were higher than the same for outpatient and inpatient departments. This is because the operating theatre and the labour room require expensive equipment, consumables and specialized personnel. Our result is consistent with the recent hospital cost study results in India which also showed that the unit costs of the operating theatres of all types of hospitals were much higher than the unit costs of several other departments such as inpatient, outpatient and emergency departments [14]. The study hospital’s material cost dominated total operating costs. This is also similar to the findings of the other hospital cost study in India which showed that for private hospitals materials cost dominated when land cost was excluded from total operating costs of the hospitals [14]. However, in this study, the materials cost was 75 % of total operating cost—much higher than reported by the other study (about 40 %)—because this study hospital was unable to provide building cost data. The consequent underestimation of the capital cost leads to a high percentage of materials cost. We tried to calculate the capital cost including building cost based on the results of the other study [14]. In that study, building cost was the highest cost component in total capital cost for the private hospital (52 %) [14], which, in turn, was 23 % of total operating cost of the hospital (authors’ calculation). If we consider the same proportion of building cost for various cost centres in the current study hospital, the increases in total capital costs will lead to 5–23 % increase in unit costs of various cost centres. The lowest increase in unit costs (5 %) will be for the operating theatre and labour room as these two occupy less floor area in total floor area of the hospital as compared to outpatient or inpatient departments (authors’ calculation based on the data of reference 14).
In the study hospital, a total of 2643 cases were admitted for surgical procedures during April 2012–March 2013 and the interventions averted 9401 DALYs using 2010 GBD study weights resulting in a cost per DALY averted of USD 165. This is much lower than the common thresholds of cost-effectiveness—an intervention is “very cost-effective” if the cost-effectiveness ratio is lower than the gross national income (GNI) per capita of the country (USD 1508–INR 91,988, for India in 2012) and “cost-effective” if the ratio is less than three times that figure [23]. As the disability weights used in GBD studies have been criticised since their inception [24–26], we did another set of calculations using 1990 GBD weights and found that the surgical interventions in the study hospital averted 2116 more DALYs using 1990 weights resulted even lower cost per DALY averted—USD 135.
Study limitations
Even though we followed the same method for cost and DALY calculations as used by several other researchers, allowing comparisons with other studies [5, 6, 14, 27, 28], the following limitations of our study merit comment. First, our cost estimates did not cover the building and land cost, causing an underestimation of total operating cost of the hospital. However, building construction or maintenance cost during the study period had been considered as recurrent cost. If we assume that the proportion of building cost in total capital cost is approximately the same as reported in other hospital cost study in India [14], the cost per DALY averted ratios for majority of the conditions will still be within the cost-effectiveness thresholds for India. Second, because detailed data of operating hours for different procedures were not available, we adjusted the operating theatre cost by the number of cases, rather than the preferred measure of operating hours. Using operating hours would change the costs of the interventions but probably not the cost per DALY averted rankings of the procedures, since in our study, most of these ratios are well below the standard threshold. Third, disability weights for some diagnoses were missing from the disease burden study tables, so we used the closest possible weights, introducing a significant amount of subjectivity. Similarly, because of the hospital coding system, some fractures were coded as humerus/tibia, and since we were not sure about the actual procedure, we used the humerus weights, which are less than for tibia in both burden of disease studies. The number of humerus fractures is therefore high in our sample, even though tibia fracture is more common in road traffic injuries in India, but this most likely reflects a coding issue, not an epidemiological orthopaedic abnormality. Fourth, the global burden of disease methods in general and disability weights in particular are based on conditions, not procedures. The hospital’s coding required us to presume the underlying condition and attribute the most appropriate weight. This bias was compounded by the fact that most of the time, identical diagnoses are not present in disability weight set of 2010 GBD study, and a ‘next-closest best’ approach was necessary. Finally, the scoring system we used for treatment effectiveness and disease severity were also used by other researchers and is internally validated but still awaiting external validation.
The present study indicates that cost per DALY averted by the surgical unit of the hospital compares favourably with several other public health interventions. As the study is based on a small private hospital in India which is not a representative hospital, the results of this study will probably cannot be used for policy purposes, however, the study result will encourage researchers to conduct large scale study in a representative sample of hospitals to establish that surgical interventions could be cost-effective and hence should be given priority in public health policies in developing countries.