Cost of Care and Impact on Quality of Life of Upper Urinary Tract Infections in South India with a Focus on Diabetics and Extended-Spectrum Beta-Lactam Producing Organisms

Abstract

Background and Objectives: 

Upper urinary tract infections (UTIs) that require in-patient care can be expensive. Comorbid conditions such as diabetes as well as UTI due to extended-spectrum beta-lactamase (ESBL) producing bacteria may affect costs. The quality of life of patients with this condition has not been described.

Methods: 

This was a cost of illness study that prospectively evaluated patients admitted with upper UTI to a medical ward in a tertiary care hospital. Direct medical and nonmedical costs, indirect costs were collected to make the total cost per admission. Quality of life was assessed using the World Health Organization Quality of Life-BREF score. We also compared costs between those with and without diabetes or ESBL infection.

Results: 

Between March 2016 and July 2017, 92 eligible patients were included in the study. The average age was 55.8 years; two thirds were diabetics. The mean overall cost of a single admission for upper UTI was INR.88, 330.2 (1370.4 USD). This was INR.96, 193.0 (1492.6 USD) and INR.1, 03,154.9 (1600.4 USD) among those with diabetes mellitus and ESBL infection, respectively. The cost was higher among those with diabetes and ESBL than those without; this difference reached statistically significance for the ESBL group. The quality of life was affected in all domains; the psychological being most affected among diabetics and ESBL infected.

Conclusions: 

The mean total cost of admission for an upper UTI in a tertiary care hospital in South India was INR 88,330 (1370.4 USD). This is higher if the patient has diabetes or ESBL organism causing the UTI. Quality of life is clearly reduced especially in the psychological domain.

Introduction

The estimated incidence of urinary tract infections (UTI) is 18 per 1000 persons per year with 58% of these being community acquired. The burden of upper UTI (Pyelonephritis [PN]) annually is estimated at 12–13 outpatient cases per 10,000 population and 3–4 inpatient cases per 10,000 population among women with lower numbers among men.[¹] PN accounts to 15%–20% of patients with community-acquired bacteremia and sepsis and 35% in a hospital setting.[²]

The estimated global cost of diabetes mellitus (DM) from 184 countries was USD 1.31 trillion or 1.8%–1.9% of global gross domestic product (GDP).[³] The indirect costs contributed 34.7%. In the low- and middle-income countries, annual direct costs range from USD 242 to USD 4129.[⁴] The prevalence in India of DM among the age group 20–79 was 8.7% (7.0%–10.6%). The total cost of DM in percent of GDP is 1.04 (0.87–1.25). Of this the direct costs account to 0.46% (0.39%–0.56%).[³] Direct costs per patient was 138.52 USD (INR.9071.67). India's estimated annual wages are 1203.54 USD (INR.78, 819.83).

A low income Indian family would roughly spend around 25% of the total family income to diabetic care. The largest expenditure for diabetes care in India is inpatient hospital admission.[⁵] The cost of diabetes in India study showed that 55% of direct costs were due to hospitalization.[⁶]

Diabetic patients have a 2–4 fold increase in bacteriuria compared to nondiabetic patients increasing the chances of an ascending infection.[⁷] The probability of having hospitalization due to PN is higher among diabetics.[⁸]

In Germany, the direct mean resource-based costs were EUR 315.90 (388.6 USD) per UTI event. Older age, higher comorbidity status, at least one previous non-UTI infection, and poorer renal function were associated with higher costs.[⁹] UTI was associated with a total all-cause incremental cost of USD 7045 (95% confidence interval [CI]: 4130, 13,051) per patient with UTI per year in the US.[¹⁰]

Recurrence rates of UTI are 25%–42% in diabetic patients which are significantly higher than patients without diabetes. The incidence of emphysematous PN and renal abscess is higher in patients with diabetes.

Bacteria that produce extended spectrum beta-lactamases (ESBLs) are also known to cause more serious UTI and could lead to increased costs.

Health is a state of complete physical, mental, and social wellbeing not just the absence of disease.[¹¹] However, the psychological and social aspects are often not measured. A tool used to measure these domains, was developed by the World Health Organization.[¹²] The relationship between cost of care and quality of life has been studied and appears varied.[¹³]

As costs and quality of life for patients with PN with and without DM or ESBL, bacteria have not been well described in India we undertook this study. The study hypothesis was that cost of care for PN would be significant with relation to paying capacity and that quality of life would be significantly affected.

The aim of the study was to estimate the cost of care in acute PN in patients. Our objectives included estimating the total cost of a single admission for acute PN, estimating the impact of diabetes as well as ESBL infection on these costs. The intangible cost (quality of life) associated was also studied.

Methods

Ethical approval

This study was conducted after obtaining permission from the Institutional review board (IRB Number 9912 dated February 05, 2016) before commencement of the study.

Study design

Setting

The cost of PN in Type-2 diabetes study is a prospective observational economic study. This was conducted among patients admitted in the general medical wards in a tertiary care hospital in South India, primarily catering to middle- and low-income groups. Participants were recruited between March 2016 and July 2017.

Participants

We included adults (>18 years), with a clinical and laboratory confirmed diagnosis of PN. They were eligible if they had fever, lower urinary tract symptoms, renal angle tenderness, pyuria and urine culture and/or blood culture growing an uropathogenic organism. We excluded subjects who had previous urological procedures, nosocomial, or catheter-associated infections and who refused consent. As we were interested in costs, we included participants from Southern states as costs such as indirect costs vary in large amount across India.

Procedure

Data were collected on a study specific clinical research form (CRF) and included risk factors including DM, chronic kidney disease, heart-failure, urine analysis, blood culture, and urine culture reports. The initial empiric antibiotic chosen, antibiotics after the susceptibility of the isolate, complications and surgical or radiology guided interventional procedures were noted.

Economic data

Patient's occupation, education, and family income per month were collected; they were stratifed into respective socio-economic class based on the Modified Kuppuswamy Score. The total duration of hospital stay and the duration of antibiotics were also noted. The quality of life was assessed by the World Health Organization quality of life-BREF (WHOQOL-BREF) questionnaire.

The abbreviated WHOQOL-BREF has four domains: physical, psychological, social, and environmental. Each domain has a “raw score” which is later converted into a “transformed score” on a scale of 0–100. This has been used in Indian studies done at New Delhi (n = 1456) and Chennai (n = 420).[¹⁴] The domain scores adjusted by age and sex, range between 14.4 and 16.2, 14.2 and 15.4, 13.9 and 14.8 and 12.1 and 14.8 for the physical, psychological, social, and environmental domains, respectively.

We used a societal perspective on costs to include both the health care costs and patient-related costs. We used the consolidated health economic evaluation reporting standard guidelines to report the data. The participants were then interviewed twice regarding costs, once at admission and another through a telephonic interview at the completion of therapy. The direct medical costs included bed charges, charge for laboratory investigations, invasive and noninvasive tests, drugs and professional charges. The direct nonmedical costs included expenditure for travel of the patient and the relatives, food during hospital stay for the patient and relative and accommodation for the relatives. In-direct medical costs included loss of wages for the patient and the relative during the time of illness. Our time horizon was only the year of the study and as we did not have costs beyond this time period we did not discount costs.

Outcomes

The primary outcome (dependent variable) was the total cost of illness for an admission for PN which is the direct medical, nonmedical cost and the indirect costs.

The secondary outcomes included comparison of costs among participants with and without ESBL (independent variable) bacterial PN as well as those with versus without DM (independent variable). We also studied the clinical outcomes such as cure, condition at discharge, and quality of life (dependent variables).

Statistical analysis

Sample size

Using an estimated cost per PN to be INR. 40,000 (620.6 USD); standard deviation (SD) INR 4000 (62.1 USD; precision (d) 1000 (15.5 USD) the sample size was calculated to be 64.

Sample Size = ([Z_α]² × SD²)/d² = 2² × 4000²/1000² = 64

Assuming loss to follow up, a sample size of 75 was taken.

All data from the CRF were entered in Epidata 3.1 software and exported for analysis to SPSS Statistics for Windows, Version 17.0. (Chicago: SPSS Inc.).

Data analysis

The baseline and sociodemographic variables that were normally distributed were described using mean and SD while nonnormally distributed data by median and inter-quartile range. As the outcome variable cost was (skewed), we used generalized linear regression with log link and gamma distribution as well as box-cox transformation. The cost comparison was done using regression analyses adjusting for age, gender, and ischemic heart disease and the difference between groups was back translated and presented in original scale of rupees. All data were reported with 95% CIs. Differences were considered statistically significant at P < 0.05.

Funding and approval

The study was funded by an internal grant of the Hospital called the fluid research grant.

Results

Between March 2016 and July 2017, eligible patients were screened and 92 patients were included in the study [Figure 1].

Figure 1: Strobe figure

The average age was 55.8 years; two thirds were diabetics with a mean 12 days hospital stay [Table 1].

Table 1: Baseline characteristics of all patients

One third each was from lower middle and upper lower socio-economic status by the Kuppusamy scale. Monthly family income [https://links.lww.com/CJHR/A4] shows a similar picture.

Escherichia coli was the most common organism isolated (68.8%) followed by Enterococcus species (6.8%) and Klebsiella species (4.5%). Of the uropathogens, 53.3% isolated were producing ESBLs.

The mean age of the diabetics was 10 years more than the nondiabetics; all patients who had hypotension were diabetics. Diabetics had more comorbid conditions such as hypertension and ischemic heart disease.

Comparing PN caused by ESBL versus non-ESBL organisms, both groups had comparable age groups and diabetes. The ESBL group had a worse renal function at presentation. The duration of intravenous antibiotics was longer in the ESBL group.

The total mean cost of a diabetic [Table 2] with PN was 18,000 rupees more than nondiabetic, although not statistically significant. The direct medical cost accounted for more than 90% of the overall costs in both groups, while direct nonmedical costs accounted for 7% in the nondiabetics group and 5% in the diabetic group. Indirect medical costs were around 1% in the nondiabetic group and <0.5% in the diabetic group. After adjusting for confounders, the difference in total cost was 25,163 INR which was nonsignificant [Tables 3a and b]. The total mean cost for an admission for PN caused by an ESBL was nearly 30,000 rupees more than non ESBL PN [Table 4]; most of the cost was direct medical cost. The direct medical cost was INR.97607.6 (1514.4 USD) compared to INR.50638.3 (785.7 USD) respectively, the difference also statistically even after adjusting for age, gender and ischemic heart disease [Tables 3a and b]. The direct medical cost contributed to around 90% in both groups, while direct nonmedical costs accounted to 10% in the Non-ESBL group and 5% in the ESBL group. The indirect cost was around 2% in the Non-ESBL group and <0.5% in the ESBL group.

Table 2: Cost description of cost in diabetics versus nondiabetics (INR)

Table 3a: Cost analysis using generalized linear regression (gamma) (INR)

Table 3b: Box-cox univariate and multivariate regression analysis for cost data (INR)

Table 4: Cost description of cost in extended spectrum beta-lactamase versus nonextended spectrum beta-lactamase (INR)

The four way cost comparison [https://links.lww.com/CJHR/A5] between diabetics and ESBL showed that diabetics with ESBL had the highest mean cost with INR.116367.7 (1805.7 USD) while nondiabetics without ESBL had the lowest mean cost with INR.64686.0 (1003.6 USD). However, the difference by the analysis of variance was not statistically significant.

Complications

Six (6.5%) patients had emphysematous PN while 15 (16.3%) developed renal abscess. Five (5.4%) required renal replacement therapy. Among them, 5/6 (83.3%) with emphysematous PN, 12/15 (80%) who developed renal abscess, and 4/5 (80%) requiring renal replacement therapy were diabetics. Four out of six (66%) with emphysematous PN, 10/15 (71.4%) who developed renal abscesses and four (66.6%) who required renal replacement therapy had an ESBL PN.

Quality of life

The overall quality of life was <50 on the transformed scale on all domains [Table 5 and Figure 2]. The psychological health was the most affected.

Table 5: Baseline quality of life based on WHO-QOL BREf questionnaire

Figure 2: Overall quality of life

Comparing the diabetics and the nondiabetics [Table 6 and Figure 3], the diabetics had a poorer quality of life and the psychological domain was the most affected. Similar results were seen among those with ESBL PN [https://links.lww.com/CJHR/A6 and https://links.lww.com/CJHR/A7].

Table 6: Quality of life between diabetics and nondiabetics

Figure 3: Quality of life in diabetics versus nondiabetics

Discussion

Medical illness in India is known to significantly affect the socioeconomic health of both the patient and family.[¹⁵] Catastrophic health expenses are defined as 40% of family income or 10% of household expenditure.[¹⁶] On an average, 4.8% of total household consumption expenditure is spent on out-of-pocket healthcare payments.[¹⁷] The proportion of households that reported catastrophic health expenditure in the NSS survey 2009–2010 was 3.5% (3.3%–3.7%).[¹⁸]

The mean overall cost of a single admission for PN was INR.88,330.2 (1370.4 USD). The cost of care for diabetics with PN was more than 96,000 INR, which was 25,000 INR more than nondiabetics after adjusting for baseline differences. Similarly, the cost for treating PN caused by ESBL producing organism, INR.1,03,154.9 (1600.6 USD) is significantly higher than a non-ESBL producing organism.

India's per capita income was INR.1, 10, 023.20 (US$ 1680 in 2016).[¹⁹] The monthly per capita expenditure in India is INR.2630 (40.8 USD) in the urban area and INR.1430 (22.2 USD) in rural India.[²⁰] Hence, an admission for PN is a catastrophic health expenditure.

The predominant contributor to the overall cost was the direct medical cost which included bed charges, lab investigations, invasive and noninvasive tests, drugs, and professional charges.

A survey in the US in 2005 found the annual societal cost for treatment of PN was $US 2.14 billion and inpatient cost for a single patient was $US 8315.02.[²¹] A study of complicated UTI found the cost to be EUR 5700 (7011.4 USD).[²²]

Data from the medicare diagnosis-related reimbursement group in the US in 1999 reported estimated inpatient costs of $US 2381.28 without comorbidities and $US 3352.65 with comorbidities.[²³] The cost of treating an ESBL organism producing PN was $US3658 (INR.239562.42) in the US.[²⁴]

The direct nonmedical costs which are often overlooked also contributed to the overall costs. This was only a small percentage of overall expenses. We noted that our patients though hailing from faraway places do not hire a place to stay to cut down on the costs.

We also noted that the indirect medical cost was small in relation to direct costs. Many of our patients from the lower socio-economic group had a caregiver who was not gainfully employed to decrease the loss of wages. The patients who had lost wages were below the 25^th centile. In the data from the US, however, the indirect costs were around 35% of the overall cost.[²⁵] In 1995, indirect costs were $US 936.

Quality of life

One of the strengths of the study was that we attempted to quantify the intangible losses in the quality of life. The psychological domain was scored lower than the social, physical health, and the environment domains. Social domain which included social relations and social support was lower than the normal South Indian population.[¹⁴] Diabetic patients in a primary care center in Kerala state that had domain-wise median scores of 53.57,58.33, 66.6, and 62.5 in the physical, psychological, social, and environmental domains, respectively.[²⁶] Elderly subjects in urban Puducherry (UT) also found most scores above 50 in all domains except in social relationships which were in the high 30s.[²⁷] None of our patients had scores above 50. Perhaps, as our subjects were admitted with a serious illness this could explain the lower scores on all domains compared to other subjects either from a general population or diabetic outpatients.

However, a study of women with UTI in Saudi Arabia found lower health-related quality of life.[²⁸] Among patients admitted with serious illness, the physical domain seemed most affected.[²⁹]

Between diabetics and nondiabetics, the psychological domain scored lower in the diabetics group and ESBL group. This could mean that the patient would take longer time to recover mentally and possible productivity loss. Overall, this intangible cost of this effect of illness needs serious consideration.

Limitations

Since ours is a tertiary care institution, the patients admitted may not represent the patient in the community developing PN, a possible referral bias. The possible under reporting of wages as has been described in economic studies, as well as quantifying the economic loss of house-wives who care for their relatives could have resulted in the low indirect costs.

Conclusions

The mean overall cost of a single admission for PN was INR.88,330.2 (1370.4 USD). The mean overall cost of an admission INR.96,193.0 (1492.4 USD) and INR.1,03,154.9 (1600.5 USD) among those with DM and ESBL infection, respectively. Those with DM and ESBL have INR 25 to 27,000 (387.9–418.9 USD) higher costs than those without, respectively. This was statistically significant for the ESBL PN group. The quality of life is lower for all patients as compared to the general population. The psychological component is especially affected in both diabetics and those with ESBL infections as compared to those without.

Contributorship statement

ED and TD were involved with planning, conduct, analysis and reporting of the work; SSRM, AJM, SS and SGH were involved in planning and data acquisition; RDS was involved in planning data acquisition and reporting of the work. BY, LJ and MB were involved in data analysis and interpretation. ED and TD are responsible for the overall content as guarantors.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Acknowledgement

We acknowledge the patients who willingly gave their time to answer questions on the cost of care and their quality of life.