1. Introduction
Diabetes is among the most common non-communicable diseases in emerging and developed nations. Substantial literature indicates the high prevalence of diabetes in low- and middle-income countries [
1]. As per the International Diabetic Foundation (IDF), it has also been found that approximately 537 million people have diabetes, and it is expected to increase by 643 million in 2030 [
2]. In addition, study indicated that 75% of people between the ages of 39 to 70 years and living in middle and low-income countries have higher risk of diabetes. In Pakistan, it was found that 6.9% of adults have diabetes, and it is expected to increase by 8.2% in 2040 [
3,
4,
5]. However, there are increased healthcare expenditures and higher costs associated with diabetes management. It has been reported that a significant portion of the diabetes cost is associated with complications and macro-vascular disease that frequently occur due to type 2 diabetes [
6,
7]. According to the Asia Development Bank (ADB), the poverty rate in south Asian countries was 15.09%, and the total extremely poor population was 256.24 million people. Furthermore, the poverty rate in Pakistan was 7.93%, with 10.26 million people living in extreme poverty. [
8]. Additionally, in Pakistan there are deficiencies in healthcare system, and there is no universal health coverage (UHC) for the population. Therefore, the increasing cost of diabetes has become a global concern in many nations, especially in low-income countries [
9]. However, this rise in the number of diabetes has increased healthcare costs, increasing productivity loss and burden on the economy. Some previous studies have emphasized that the high cost of type 2 diabetes in healthcare organizations has increased the overall loss of economic activities and productivity [
9,
10,
11,
12,
13].
The management of diabetes has gained international attention, and several countries—including Saudi Arabia (21%), Sri Lanka (16%), Malaysia (16%), Mexico (15%), and the US (16%)—spend a significant amount of their health expenditures on diabetes care [
14]. The American Diabetes Association’s cost estimates for the US population are notable instances of cost-of-illness research; the total burden for 2012 was projected to be USD 245 billion, taking into account higher health spending and productivity losses caused by diabetes. The relevance of labor market consequences for the overall economic burden was highlighted by the fact that indirect expenses accounted for 28% of the total costs of diabetes in this research [
15].
There has only been a few cost of illness studies being conducted in Pakistan. The study by Khowaja et al. reported the direct cost of diabetes management at USD 197 in 2007. It was also estimated that medication and laboratory investigations contributed to the highest direct cost incurred at 46% and 32%, respectively [
13]. Another study reported in 2018 that the total direct cost of diabetes management is USD 332, with 60.4% accounting for the total medication cost of management. In these studies, there was a lack of indirect cost computation; therefore, no indirect cost data are available for the region [
4]. A study in Bangladesh reported that the overall cost of diabetes management was USD 864.7, with 60.7% accounting for the direct medication cost [
16].
To be sure, developed nations are not the only ones where there is evidence of substantial economic consequences. For instance, Seuring and colleagues examined the findings of 86 cost-of-illness studies published between 2001 and 2014 and discovered evidence of a significant economic burden in low-income and middle-income countries (LMICs), with annual direct costs ranging from USD 242 to 4129; (2011 purchasing power parity) per capita and indirect costs ranging from USD 45 to 16914 per capita. It is challenging to compare cost-of-illness studies within and between nations, since the costs included in the calculations and the methodologies used to quantify costs differ significantly in the literature in circulation [
17].
Studies on the “cost of illness” (COI) explore the effects of a disease on people, communities, and the nation from various angles. COI research aims to identify and estimate all direct, indirect, and intangible costs associated with a specific condition. The output estimates the societal financial costs associated with a particular disease. It is widely acknowledged that determining the total financial burden of a disease is a valuable tool for developing national and international health policies [
18].
Approximately 12% of healthcare expenditures are attributable to diabetes management, which is about USD 827 billion globally [
19]. Most studies have highlighted a substantial increase in the number of people living with diabetes, along with the rise in poverty [
20,
21,
22,
23,
24]. Moreover, this is coupled with the fact that medicine prices have increased substantially in Pakistan over the past two years [
25,
26]. Therefore, the current study aimed to assess the cost of illness of diabetes. Pakistan is a lower-middle-income country with limited healthcare resources, and this study will help the healthcare regulatory bodies and policymakers to optimally allocate resources to manage diabetes.
2. Materials and Methods
2.1. Study Design and Setting
The current multi-center study assessed the direct and indirect healthcare costs of patients with type 2 diabetes (T2D) in Pakistan. The study was conducted between October 2021–February 2022. The study setting was outpatient clinics dealing with diabetes.
2.2. Study Population and Sampling
The study population was patients diagnosed with T2D. T2D patients were included if their age was ≥ 18 years, did not have any malignancies, and did not have other chronic diseases or drug abuse. The prevalence-based sample size was calculated for the current study. The sample size was calculated using OPENepi online sample size calculation software. A sample size of 1782 ≈ 1800 was calculated based on the 12% prevalence of T2D [
26], with 20% relative precision, an odd ratio of 1.8, and 80% power [
27].
Pakistan is a lower-middle-income country comprised of four provinces, Punjab, Sindh, Baluchistan, and Khyber Pakhtunkhwa, and the sample size was distributed equally (each = 450). Three districts were selected randomly from each province, and the respondents were recruited equally from each section (each district = 150 participants). The study participants were recruited through cluster sampling, and three clusters from each district were collected (50 participants per cluster). All metropolitan cities in Pakistan were randomly selected (including rural and urban settings) within each province were included in the study sample, which was chosen using a stratified two-stage cluster design. The patients were recruited by seven teams of trained field workers overseen by three project leaders. The field team gathered basic demographic information and asked the selected sample’s consent to provide pertinent facts and answer all the questions about the resources utilized.
2.3. Study Tool and Data Collection
A self-developed structured questionnaire was used for data collection (
Supplementary Material File S1). The developed questionnaire was incorporated into an online server using google documents, the secured web-based application used for data collection. Before the main study, a pilot study was conducted with 35 participants to assess the internal consistency of the questionnaire. A Cronbach alpha value of 0.87 was obtained, indicating a good internal consistency. After obtaining informed consent from patients selected randomly in the diabetes outpatient clinics, patients were invited to complete the structured questionnaire online (
Figure 1).
The questionnaire comprised of six different sections.
Section 1 comprised personal identifiers (e.g., which hospital the data were collected from).
Section 2 comprised sociodemographic and general information.
Section 3 contained health status and behavior information and
Section 4 with co-morbidities/complications and treatment information.
Section 5 and
Section 6 comprised treatment regime information and diabetes cost information, respectively.
The patient’s medical records from the clinics were also reviewed to retrieve the patients’ clinical records and the number of services received throughout the preceding year from the recruitment date. Moreover, the current prescribed medicine data were obtained for the prescription. The data were collected using an electronic database, and further information was added to the final database.
2.4. Ethical Approval and Consent to Participate
The Bio-Ethics Committee (BEC) of Bahauddin Zakariya University, Multan, permitted this prospective observational study in Tertiary Care Hospitals & Diabetes Outpatient Clinics (ACAD/EXT/01/2022).
Written informed consent was taken from all the participants included in the study without assistance from others; in the case of illiterate individuals, consent was obtained from their legal guardians. The study was conducted as per the Declaration of Helsinki. The current study was reported according to “The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)” guidelines.
2.5. Calculation of Cost
2.5.1. Direct Medical and Non-Medical Cost
The direct cost was premeditated using a bottom-up approach. It was divided into two categories: direct medical costs, which include the costs of hospitalization, out-patient visits, medicine, laboratory tests, and other service utilization (including the use of self-monitoring blood glucose and consumables); and direct non-medical costs, which include the costs of transportation, hotel stays (as hospitals are located far from home), and meals (special diet plan, as recommended by international diabetes federation).
To compute the direct cost, the micro-costing approach was utilized to identify cost items in as much detail as possible. Each medicine and therapy’s cost per day was calculated. The costs of treatment, consulting, and laboratory tests that patients paid out of pocket (OPP) in the selected regions were obtained from a tertiary-level hospital. Private health insurance is limited in Pakistan. Hence, patients were responsible for all costs. Individually, participants’ replies to the questionnaire were used as a reference for other components of direct medical and non-medical costs. The unit cost was multiplied by the number of medical services supplied the previous year to determine each direct cost component. The hospital bill for each patient was calculated using their hospital record, which the hospital’s accounts department provided. All direct medical and non-medical expenditures were added to the total direct cost.
2.5.2. Indirect Cost
The indirect cost for patients and their caregivers was calculated on the way to the hospital. The productivity time lost between out-patient visits and during hospital admission was documented based on information provided by patients and their caretakers. The indirect cost for productive workers in the formal workforce or housewives was calculated using the human capital approach but not for people who were unable to work (retired or ill-health) or who chose not to work. The productivity loss of housewives was calculated using the minimum wage rate of PKR 120 (USD 0.7) (USD 2484/annum) set by the court, as well as the median income of the participants in the formal workforce. Total direct and indirect costs were added together to determine the total cost.
All expenditures were computed in Pakistani rupees (PKR), then converted to USD using the mid-year currency conversion rate for 2022 (USD 1 = PKR 180).
2.5.3. Data Analysis
The data from the online database were imported into Microsoft Excel. The data were thoroughly checked for errors and then exported into SPSS. Data were analyzed using SPSS v25. The categorical data were presented as frequency and percentages, while the continuous data were tabulated in mean (standard deviation, SD) and median. Using inferential statistics, the study variables’ significance was established (Mann–Whitney U test and Kruskal–Wallis tests). Pearson correlation coefficient analyzed the relationship between variables. A median regression analysis was implied to evaluate the contributing variables to the direct and indirect costs. The 95% confidence interval (CI) of regression coefficients was determined by bootstrapping. The p-value ≤ 0.05 was considered significant.
4. Discussion
COI analyses may provide helpful information for decision makers for distributing limited resources or desired amenities. Previous studies conducted in Pakistan were region-specific, and the cost component did not collect hospitalization data to evaluate the economic impact of diabetes care. These studies were limited to Karachi and South Punjab and did not provide enough information about the entire country. This study aimed to evaluate the nationwide economic consequences of diabetes. All attempts have been made in this study to assess all plausible cost components. Policymakers are interested in maintaining or lowering COI; hence, this study was primarily done to determine the cost of diabetes care in Pakistan. Individuals, families, and governments must allocate resources for health care based on existing evidence while also addressing other basic needs.
The current study was focused on the direct and indirect costs of diabetes management. In the present study, most patients were on oral anti-hyperglycemic drugs. The annual total cost of diabetes care was USD 740.1, amongst which the share of the direct cost was USD 646.4, and the indirect cost was USD93.65. The medicine (USD 274.5) and hospitalization (USD 319.7) shared the most direct cost. At the same time, the productivity loss of the patients had the highest contribution to the indirect cost (USD 81.36).
The calculation of diabetes-related costs has gained importance as the prevalence of this disorder increases with every passing day. In Pakistan, where healthcare resources are limited, it is imperative to calculate diabetes-related costs accurately. This will aid in the efficient allocation of healthcare resources. Direct costs are frequently higher than indirect costs, according to a thorough assessment published in 2015. Annual direct expenses ranged from USD 242 for an out-of-pocket study in Mexico to USD 11,917 for a diabetic cost study in the United States.
In contrast, indirect costs ranged from USD 45 in Pakistan to USD 16,914 in the Bahamas. In contrast to higher-income countries (HICs), patients in lower-middle-income countries (LMICs) bear a considerable financial burden through out-of-pocket medical payments. In a 2018 systematic analysis, the annual average cost of disease for T2D in low- and middle-income countries was estimated to be between USD 29.91 to USD 237.38. This difference may be affected by the different methods used in various studies [
28]. The study reported from Bangladesh estimated the average annual cost of diabetes management was USD 864.7 per patient, which was in line with our study findings [
29].
The other key finding of this study was that the patients hospitalized during the past year due to any complication related to diabetes incurred three times higher costs than those without hospitalization. The direct cost for the patient hospitalized due to diabetes-related complications was USD 1038. At the same time, medication and hospitalization contributed the most to the direct cost for those patients, with a mean cost of USD 432 and USD 419.4, respectively.
With a current diabetes prevalence of 16.98% and a total population of 220.9 million in Pakistan in 2020, the predicted direct cost of diabetes is 495.0 billion PKR, accounting for roughly 73.7 percent of Pakistan’s total federal and provincial yearly health expenditure (671.4 billion PKR). According to our research, diabetes treatment accounted for 23% of household earnings in the lowest-income groups. These findings are comparable to an Indian study, in which low-income diabetic individuals spent 25% of their monthly family income on diabetes care. Diabetes care is a critical impact on patients’ stress and the financial burden on the patients’ families and the government. Regarding socio-demographic characteristics, older patients, rural inhabitants, and those with a high socioeconomic status SES showed a significant link with direct diabetic expenditures. A prior study in Pakistan found a non-significant difference between age and direct medical costs [
4]. Previous research from India and Singapore contradicted our findings; those investigations revealed no statistically significant differences between age and direct cost [
30,
31].
Regarding clinical parameters, direct costs increased dramatically as the disease duration and co-morbidities increased. Patients with one and more than one co-morbidity had 34.2 percent and 39.7 percent higher mean direct expenses, respectively, than patients without co-morbidities (
p < 0.01). These findings are comparable to a cocoa study, which observed that a 91.8 percent increase in costs for patients with one or more co-morbidities (
p ≤ 0.002) [
32]. In both emerging and developed countries, similar tendencies have been observed.
In the current study, it was found that direct cost was the most common cost in the total cost of diabetes care. The share of the medicine was highly accounted for in the direct cost. The study reported from Kenya showed similar findings, where the treatment accounted for 52.4% of the total direct cost of diabetes care [
33]. Similar results were reported from Bangladesh, where medicine contributed 60.7% of the total cost [
29]. Moreover, in the current study, it was observed that the productivity loss of the patient was mostly accounted for by indirect cost. However, it was observed in the published literature that absenteeism, presenteeism, and premature mortality significantly contributed to indirect cost [
34,
35,
36]. In addition, it was also noted that food and travel significantly contributed to the indirect cost [
35].
A significant correlation was found between the socio-demographic parameter and their association with the overall cost of diabetes. The higher cost was reported in males, <60 years of the study population, with the highest cost of USD 1376.9. The study reported a higher cost associated with prolonged diabetes duration and higher HbA1c. Previous literature, considerably in low and middle-income countries, reported similar findings. It was reported that the higher cost of diabetes management was associated with the higher prevalence of comorbid conditions [
37,
38,
39].
According to world bank data, the per capita GDP of Pakistan was USD 1193.73, as reported in 2020. The current study reported that the overall diabetes management cost was 62% of the per capita GDP of Pakistan. At the same time, Pakistan’s GDP per capita healthcare spending is at 3.38%. However, with such scarce spending on healthcare and the disease burden of diabetes in the region, it is high time for policymakers to consider universal health insurance coverage to minimize the overall diabetes management burden on the population [
40,
41].
Considering the overall cost of diabetes management to Pakistan, the population-based cost of illness computed, based on the prevalence of type 2 diabetes mellitus reported by the International Diabetes Federation of 33 million people [
9], the total burden of diabetes management is USD 24.42 billion. The total burden of diabetes management is almost 1.60% of Pakistan’s GDP, as estimated from World Bank data for 2021. In the USA, the overall cost of diabetes management, estimated in 2017, was USD 327 billion. That is estimated to be 1.67% of the country’s total GDP [
42]. It was documented that medicine prices is the significant contributor to the overall diabetes management cost. In lower-middle income countries including Pakistan there are challenges related to forming medicine pricing policies which ultimately led to higher medication cost. Accessibility to medicine has been essential in enhancing patient outcomes and lowering mortality on a worldwide scale. It is crucial to remember that drug costs play a significant role in increasing access to medications. Studies and empirical evidence have demonstrated that pricing for the same brand of medications vary greatly between nations [
43].