ECONOMIC EVALUATION Out of pocket payment for diabetes mellitus in a public hospital of Karachi

Pakistan


Introduction
The World Health Organisation (WHO) report of 2011 stated that 12.9 million inhabitants of the world suffer from diabetes mellitus (DM) and the number is on a continuous rise. 1 The monetary loss due to the disease is expected to increase from United States dollar ($)113bn in 2009 to $336bn in 2030 with a cumulative output loss of $47 trillion globally over the next two decades. 2,3It is expected that there will be a staggering 69% increase in numbers of adults with DM in developing countries and a 20% increase in developed countries between the years 2010 and 2030.
In Pakistan, the non-communicable diseases (NCDs) account for around 55% of overall deaths in the country. 2 According to the report by Oxford poverty and human development initiative (OPHI) in 2017, 44.2% of population in Pakistan was poor among whom 23.7% were in severe poverty and 20.7% in destitute poverty bracket.Moreover, 15.1% were vulnerable to poverty. 4 Few state-run health insurance schemes have been introduced like Waseela-e-Sehat programme under the umbrella of Benazir Income Support Programme (BISP) having insurance coverage up to Pak rupee (PKR) 25,000 per family per annum 5 and Prime Minister's National Health Programme with Pakistan Sehat (health) Card providing insurance cover for up to PKR50,000 for inpatient services, maternity cover, fractures and injuries, and up to 250,000 for heart diseases, diabetes coverage, organ failure and cancers. 6However, the empanelled hospitals are very limited and the coverage is not available countrywide.In effect, health coverage and insurance for the low socioeconomic group is not sufficient.

Diabetes being one of the prevalent NCDs in Pakistan
showing a rising trend is not only a problem for patients, physicians and healthcare staff, but its repercussions can be heard in other sectors like finance, education, population welfare and so on.As a matter of fact, once the disease develops, the cost for its management surpasses the cost and effort used for prevention, if done at the right time.
8][9] One study was conducted specifically to assess the cost of diabetes care in an outpatient care clinic. 10However, these studies lack information regarding public-sector hospitals and are more than five years old.The current study was planned to estimate the average OOP expenditure on the DM management irrespective of the treatment regimen in use.

Subjects and Methods
The cross-sectional study was conducted from December 2016 to May 2017 at Jinnah Postgraduate Medical Centre (JPMC), Karachi, and comprised type 2 DM (T2DM) patients coming to the outpatient department (OPD).The hospital not only caters to people coming from all over Karachi but also provides care to many coming from the rural areas.After approval from the ethics review board of Jinnah Sindh Medical University (JSMU), Karachi, the sample size was calculated using WHO software 11 at an expected standard deviation (SD) of 324, as reported by a previous local study, 10 confidence level of 95% and bound on error of 35.Using non-probability convenience sampling, all diagnosed T2DM patients were approached.After written informed consent was obtained from all the enrolled subjects (Annexure-I), they were all interviewed while they waited for their turn to be consulted in a separate room.The questionnaire comprised questions in both English and Urdu languages for convenience, although the questions were mostly asked in Urdu.The data collectors had good command on both the languages since it was one of the essentials while hiring them.
OOP expenditure was defined as charges or fees for medical consultation, investigation, medication and travelling to get the treatment paid with one's own money rather than with money from another source. 12tructured questionnaire (translated into Urdu) was used to obtain detailed information on sociodemographics, occupational characteristics, T2DM history, medication history, income and expenditure history, expenditure details and expenditure source (Annexure-II).The questionnaire was adopted from relevant previous studies. 10ta collectors were medical students supervised by the principal investigator.After training, the questionnaire was piloted and revised according to feedback provided by data collectors.Data was entered in Microsoft Excel software in the form of numeric codes assigned to different variables.It was analysed using SPSS 20.Descriptive statistics were run for the socio-demographics.Frequencies and percentages were calculated for categorical data while means with SD were calculated for quantitative variables.Differences in average OOP expenditure indifferent groups was analysed using independent t-test for two groups and analysis of variance (ANOVA) for more than two groups.Non-parametric tests were used if the dependent variable did not meet the assumptions of normality.The percentage of income spent on the disease was calculated by dividing the OOP expenditure by the total income of the participant or the person sponsoring the payment.Differences for average percentage of income spent on the disease were also computed for the groups.

Results
Of the 336 subjects, 178(53%) were males and 158(47%) were females.Overall mean age was 52.20±12.64years, and 170(50.6%)were illiterate (Table -1).No statistically significant difference was found between the genders in median OOP expenditure (p=0.335).OOP expenditure was significantly higher among the more educated (p<0.001), the higher income group (p=0.020) and those with longer duration of disease (p<0.001 The females spent more of the monthly household income on T2DM management than males (p=0.016).The percentage of income spent increased with the level of education (p=<0.001).Large family size meant less percentage of income spent on the disease (p=0.009).People earning PKR15,000 or less per month spent higher percentage of income on the disease compared to people earning >PKR15,000 per month (p<0.001).
The percentage spent also increased with duration of the disease (p=0.019)(Table -3).
Overall, the mean total OOP expenditure on diabetes management was PKR 2227.11±2217.70per month.Major heads of expenditure included medicines, consultation, transport and laboratory investigations (Figure -1).
Besides, 220(65.48%)participants were dependent on others to pay for their    treatment and 107(32%) could not cover the disease expenditure and had to borrow money.Basic necessities compromised due to OOP expenditure on T2DM management were utility bills, education and food expenses (Figure -2).

Discussion
The study found that on an average, the participants spent around PKR2227.11($21.7) per month on the management of diabetes which makes around an average of 9.7% of their monthly income.These findings show an increase over numbers reported in 2007 when the mean direct cost of diabetes was PKR965 per month. 10his increase may be explained by inflationary trends.
Another study done in the private sector in India found that around 17% of annual household income was spent on diabetes. 9Since the current study was done in a government hospital, most of the services were either subsidised or were free of cost.Moreover, it also did not cover patients of type 1 diabetes which also may have contributed to a lower percentage of income spent on the disease.
The major cost item was medicines, costing a high average of 80.82%, which is consistent with a previous study in Pakistan in which the most (46%) was spent on medicines.
These findings can be explained by the service delivery issues at public hospitals where drugs are not available to patients and they have to buy all or most of their medicines themselves.However, the cost incurred on laboratory investigations was around 4.23% which is considerably less than what was depicted in an earlier study in which about 32% of the cost went to investigations alone. 10his finding can be supported by the fact that in government hospitals, the cost of laboratory investigation is highly subsidised.Moreover, uncomplicated diabetes usually does not require extensive investigations on a frequent basis.Nevertheless, the current study found that 9% of expenditure was related to consultation.The reported spending on consultation by patients presenting to public health facility point towards parallel utilisation of the private sector by these patients possibly to obtain clinical opinion from another doctor or maybe in cases where emergency care or when opinion of a private practitioner with specific expertise is required.Moreover, relatively higher treatment expenditure was reported by the participants with education of matriculate and above, who spent around 17.89% of their monthly income on disease management.It indicates that as the level of education increased, there was an increase in awareness about the disease and affordability, leading to better compliance of taking medicines and opting for investigations.As shown in a 2007 study done in India, as the education level increased, people tended to spend more on the disease and, hence, had better compliance for management of disease. 13Furthermore, with the increase in income, people might also be going for expensive medicine and choosing private laboratories.Family size was found to be inversely proportional to expenditure on disease which can be explained by higher affordability of relatively smaller families.A study done on Mexican-Americans also supports the finding as it reported that smaller family size was associated with better self-care for diabetes. 14P expenditure was higher as duration of disease and complications increased.A study done in France reported that as duration of the disease increased, patients tended to move from single therapy to either double, triple or insulin therapy which directly increased their cost.Furthermore, duration also increased the number of comorbids, leading to as much as 86% increase in the total OOP expenditure on the disease. 15e current study found that the participants with income <PKR15,000 per month had to spend higher percentage of their income.This indicates that poor are more likely to face the economic burden of the disease.A study in India found that the poorest households tended to spend as much as seven times more than rich household members on the management of diabetes. 13Different local and global studies also substantiate these findings related to huge impact of high OOP expenditure on all socioeconomic classes among whom the poor quartile is the most affected. 9,10,16,17Moreover, insurance coverage for the poorest quartiles is much lower than relatively higher income groups.For example, in a study conducted in India, only 6.4% urban low-income group received medical reimbursement compared to 21.3% in the high income group. 13Issues related to insurance coverage for the poor are similar in Pakistan as only one-third of expenditure on health is contributed by the government and insurance, while two-third of expenditure is OOP.Few state-run health insurance schemes have been introduced, like the Waseela-e-Sehat under BISP(5) and the Pakistan Sehat Card. 6However, the empanelled hospitals (those included in the list of providers by the government) are very limited and the coverage is not available countrywide.The issues in these insurance schemes are mostly compounded by inadequate allocation of the money, late disbursements, corruption and lack of trained staff who can utilise the funds appropriately.
This calls for a mandatory insurance cover for all the socioeconomic classes with poorer classes being on high priority.
In our study, 32% participants were so poor that they had to borrow money for the treatment of diabetes.The results are almost identical to the study done in India showing people had to borrow as much as 35% for covering the cost of management of diabetes. 9Moreover, diabetes not only has a direct economic impact, but also has a huge indirect impact as it leads to loss of huge amount of workforce which is in their productive years which, in turn, further increases the economic burden on the family. 13,16The current study also explored the most compromised basic need due to expenditure on disease.Most of the participants were unable to pay the utility bills which was the most compromised need followed by educational and food expenses.Other studies have also documented indirect costs of diabetes incurring informal economic losses. 10,18e current study specifically aimed at estimating OOP expenditure on the management of T2DM among patients presenting at a public-sector, tertiary care hospital.OOP expenditure was higher among diabetics having disease <6 years of duration compared to those having it for 6-13 years; possibly due to increased co-morbids.This suggests that early detection and treatment of DM can reduce related economic burden on patients and their families.
While the study attempted to explore a wide range of aspects related to OOP expenditure on diabetes, it was conducted on a small scale at a single centre due to limited research funds.The data would be more meaningful if compared with data related to privatesector hospitals.Moreover, the study only represents urban population, which, again, is a limitation.The overall costs on diabetes may have been underestimated because type 1 diabetics were not enrolled and inpatient cost was not evaluated.Indirect costs were also not part of methodological assessment.Despite the limitations that would have underestimated the average OOP expenditure, the effects observed on even underestimated costs were quite remarkable which is an alarming sign.
The current study raised important questions on the effect of OOP expenditure on management of a chronic condition like diabetes which is often an underresearched area.The issues raised should further be investigated on a larger scale and with improved methodology for a more in-depth understanding of the economic impact of chronic diseases in Pakistan.Furthermore, policies to prevent the occurrence of DM should be put in place to lower the burden of disease in the general population so that the expenditure on disease decreases in the long run.There should also be a universal model for insurance which may cover the whole population so that the finances don't continue obscuring the care for diabetes.

Conclusion
High OOP expenditure was incurred by people diabeties.The coverage of social security nets and individual insurance should be broadened to help improve diabetes management.

Figure- 2 :
Figure-2: Prominent types of basic needs compromised due to out-of-pocket expenditure on the disease (n=101).

Table - 2
: Comparison of median out of pocket expenditure (PKR) on the disease among different groups (n=336).

Table - 3
: Comparison of Median percentage of Income spent on the disease between different groups (n=336).