Study Design and Tools
This study was a cost-consequence analysis to examine the expected health gains as well as costs of scaling-up certain interventions to reduce the burden of mental disorders in different groups of patients suffering from anxiety, depression, bipolar disorder, and epilepsy in Iran. The scaling-up scenario was thus tested against the no scaling-up scenario with the existing level of care offered in the system. The data required for the study were collected in the first quarter of the year 2020 and the time horizon was a 10-year period from 2020 to 2030. Although the amount of services provided in the private sector is considerable and includes 75% of outpatient services [11], in this study we assumed ministry of health as the single provider in order to restrict the scope of research and gather high quality data to make precise estimations. It was therefore assumed that all the services were being provided by the government within the facilities of the Ministry of Health and Medical Education (MOHME) and the analysis was fulfilled based on its perspective.
One Health Tool Module on MNS Disorders
The OneHealth Tool (OHT), a software tool developed by the international costing experts from the World Health Organization (WHO) and other United Nations (UN) agencies, was employed to start this study. The mental health module of the OHT was accordingly developed in order to ensure that the national mental health development plans have been carried out within a framework of assessment of overall health system capacity and to take financial sustainability and outcomes-based planning into account [12].
Research Planning
Consulting with three main national planners at the Department of Mental Health in the MOHME who are in charge of this program, priority MNS disorders including depression, anxiety, bipolar disorder, and epilepsy were selected.
The main criteria to select these interventions include 1) Provision of care within the centers affiliated by ministry of health: for instance, the drug abuse treatment was excluded as the services are mainly provided in other sectors. 2) Burden of diseases: as reflected in the introduction, these disorders are among most prevalent ones in national and global levels and 3) The inclusion of the interventions in OHT impact module: Since not all of the interventions are included in this module to estimate health impacts , we checked the available interventions and excluded ones which are not provided in impact modules and then we finalized eight interventions.
It is noteworthy that these interventions are recommended in WHO Mental Health Gap Action Program (mhGAP) guideline [13] as standard and effective interventions. So, their effectiveness is proved through RCTs and Systematic reviews which are formulated in the OHT to estimate the health impacts. Although national estimation of effectiveness would better describe the situation , it hasn’t been in the scope of this study to estimate these measures and no national studies were available too. Therefore, we relied on the global and regional evidence of effectiveness in the OHT.
Moreover, appropriate mental health care packages and scenarios, current and target coverage levels for specific intervention strategies, and scaling-up period were identified through consultations with the expert group of national planners and program managers.
For clinical-level consideration of resource use profiles for different disorders and interventions, unit costs and prices for health care services and commodities (such as those for staff salaries, outpatient visits, and psychotropic medications), one of the pilot sites (namely, the city of Oskoo in East Azerbaijan Province) was selected. This new service model was initially implemented in three cities [14]. Although a formal evaluation was not performed to compare the process in these three pilot sites, based on the regular observations, national managers of the program believe that Oskoo is the best practice to the program, benefiting from excellent documentation and assessments. To collect these inputs, the research team, working with local team members and other national staff, identified and utilized local data sources and visited the site in a three-day tour. The data checklist adapted from the study by Chisholm [12] was also modified and used in order to facilitate and document the process of data contextualization.
Data Contextualization
Health Impacts
The selected metric for summarizing health effects at the population level is (healthy) life years (LYs) gained (equivalent to disability-adjusted life years (DALYs) averted), where one DALY could be thought of as one YLL. (Healthy) LYs is also computed with reference to country-specific life tables that have been already built into the model, and reflect the combined time spent by the population in a particular state of health with a known degree (or free) of disability. Default disability levels in OHT were drawn from the GBD 2010 study [15]. Implementing or scaling-up an effective intervention in the population was thus modeled to reduce the time spent in a disabling state, either by reducing prevalence (e.g., by decreasing the number of new cases or increasing the remission rate), or by improving the level of functioning in people with the condition in question [12].
Epidemiology
Iran has a history in mental health surveys and it is therefore possible to make use of high-quality epidemiological data from the national surveys [16, 17, 18]. However, to meet age-stratified data sheets of the OHT on the prevalence and incidence rates for all disorders, the 2017 GBD estimates available in GBD result tool (http://ghdx.healthdata.org/gbd-results-tool) were used. The prevalence and incidence rates were extracted for the four disorders including depression, anxiety, bipolar and epilepsy and the data were sorted for male and female in 17 age groups. Then, to meet the OHT requirements for data entry, these rates were modified to cases per 1,000 population.
Estimating Resources, Costs, and Coverage Levels
The key categories of health service costs in the OHT included drug and supply costs (e.g., daily dose of a generically produced first-line anti-psychotic or anti-epileptic medication), costs of response to ambulatory contacts by mental health or general health workers (such as psychologists, counselors, and community health workers), and costs of hospital-based outpatient/inpatient care services. In addition, program-level resource needs were identified, including overall program management and administration as well as training [12].
Intervention Costing
Resources used in terms of drug and supply, staff salaries, and outpatient visits were documented through field visits to the pilot site and the associated costs were subsequently estimated based on accounting documents available in the program management office in order to estimate the average cost of each intervention per case.
As described in the introduction of the new care model, basic package of care (e.g. basic psychosocial treatment plus medication therapy of moderate cases of anxiety/depression) were being provided in health care facilities and severe cases of those disorders in addition to all bipolar cases could be referred to the CMHCs for intensive care and social support. It is noteworthy that the WHO mhGAP guideline was practiced with slight changes in real time spent and availability of recommended drugs in the whole system of integrated mental health care in PHC.
All three levels of care (health care facility, CMHC, and general hospital (psychiatric ward)) were accordingly visited and all staff including community health workers, socio-mental health experts, general practitioners, psychologists, psychiatrists, and social workers were interviewed in order to identify activities and estimate time spent for each patient and intervention. The drug and supply list used for each intervention was also documented and the price list was acquired from the health information system (HIS) of the hospital with a particular focus on most-prescribed medications available in the primary care setting. It should be noted that drug and supply prices have the same tariffs across the country in public hospitals and the PHC network.
Program Costing
Program specific costs concern the cost incurred at the national and provincial level to manage and supervise the program. Information on the program-specific staff required for scaling-up were obtained from background strategic documents of the program available at the Department of Mental Health including a unit cost study of new socio-mental program, so-called SERAJ, in the year 2018 [19]. In order to estimate the cost of supervision and auditory visits as well as necessary trainings, the district managers were interviewed in the field visit and costing parameters including transportation, trainings and allowances to take part in missions were identified.
Total Cost
The total costs of scaling-up an intervention in a given year for a country was thus derived by multiplying resource use needs by their respective unit costs to give an average cost per case. It was then multiplied by the total number of cases, expected to receive a particular intervention (given by the prevalence of the disorder multiplied by the rate of treatment coverage of specific intervention strategies in the population), that is, total cost=population×prevalence rate×coverage×treatment cost per case [12]. To estimate the costs, the expected inflation rate for the period: 2020 -2030 was applied using extrapolation based on the actual inflation rates for 2010-19. Since all the costs were calculated in Iranian Rial (IRR), to convert the costs in US$, the exchange rate for the study period was estimated through extrapolation based on the UN exchange rates for the past 10 years.
Coverage
In order to estimate present and future coverage levels, we have conducted three rounds of panel discussions with three national policy makers and three provincial managers who have been in charge of this program at national and district levels. The first panel was conducted at the beginning of the research, the second one was just at the early beginning of data collection and the last one during the filed visit in the pilot site. All the estimations of coverage were made based on the population covered in each area and expected coverage in the future based on the available budget. In order to make sure about coverage levels, we re-checked measures before and after field visits and examined different scenarios.
The estimates of the baseline coverage were based on the coverage of health care facilities per population. Given the expected challenges of scaling-up the new model at the national level, a modest coverage target of 30-40% by 2030 was set for the intensive packages of care and 60-70% target coverage was considered for the basic package, which have already had 40% baseline coverage.
Assumptions
In this study, the volume of services provided in the private sector was not estimated. It was assumed that all the services had been provided by the government within the facilities of MOHME. Besides the base case scenario which is provided below, we examined seven others scenarios in order to provide a sensivity analysis.
Currencies reported
At first, all the costing parameters were estimated in IRR (Iranian Rial) as local currency values were used in the context of ongoing policy dialogue, then all the cost values were converted into the US dollars for ease of interpretation and comparison. The exchange rate was captured from the UN exchange rates in June 2020 which equals to 194,881 IRR.
In addition, the final result of cost per each health life years were also reported in Intl dollars (PPP adjusted) to provide a more realistic view while the results are compared to that of different countries. In order to convert values to Intl dollars we used the cost conversion tool developed by Campbell Collaboration (https://eppi.ioe.ac.uk/costconversion/). The underlying methods of this tool is provided in the paper published in the journal of Evidence and Policy [20].