Healthc Inform Res Search


Healthc Inform Res > Volume 23(4); 2017 > Article
Park, Lee, and Lee: Association between Health Information Technology and Case Mix Index



Health information technology (IT) can assist healthcare providers in ordering medication and adhering to guidelines while improving communication among providers and the quality of care. However, the relationship between health IT and Case Mix Index (CMI) has not been thoroughly investigated; therefore, this study aimed to clarify this relationship.


To examine the effect of health IT on CMI, a generalized estimation equation (GEE) was applied to two years of California hospital data.


We found that IT was positively associated with CMI, indicating that increased IT adoption could lead to a higher CMI or billing though DRG up-coding. This implies that hospitals' revenue could increase around $40,000 by increasing IT investment by 10%.


The positive association between IT and CMI implies that IT adoption itself could lead to higher patient billings. Generally, a higher CMI in a hospital indicates that the hospital provides expensive services with higher coding and therefore receives more money from patients. Therefore, measures to prevent upcoding through IT systems should be implemented.

I. Introduction

Health information technology (IT) can assist providers in ordering medication and adhering to appropriate guidelines in the treatment of disease [1] while improving communication among providers within and between organizations [2]. Health IT is particularly useful for the gathering, organization and display of information, as well as facilitation of follow-up and other vital functions. [3,4]. Moreover, health IT could help improve the quality of care by reducing errors [5,6,7] and improving patient safety by preventing adverse drug events [8,9,10].
Some studies have reported that health IT adoption is associated with reduced expenses in the hospital setting due to improved efficiency after health IT adoption [11]. However, the effect of health IT on healthcare costs still is not completely understood because health IT could be used by providers to increase billing (i.e., up-coding) [11,12,13]. Recently, increasing concerns have emerged that the implementation of IT systems is likely to make it easier for providers to change patients' billing codes, and this could contribute to rising health expenditures. Some vendors explicitly advertise that their EMR will help physicians raise the level of billing codes. For example, one EMR vendor advertises on its website that its product will result in an increase of one coding level for each patient visit [14,15]. This up-coding issue related to IT adoption could damage the national healthcare financial system. According to the US Office of Inspector General (OIG), the federal government recovered nearly $5.7 billion and $3.35 billion in healthcare fraud cases in 2014 and 2015, respectively.
Despite all of these concerns, few studies have investigated the relationship between IT system adoption and healthcare providers' revenue-enhancing practices. Li [12] and Ganju et al. [13] used the hospital-level Case Mix Index (CMI) as a payment measure and found that when hospitals adopted health IT systems including Electronic Medical Record (EMR) and computerized physician order entry (CPOE) functions, reimbursement was inflated through diagnosis-related group (DRG) up-coding. However, Adler-Milstein and Jha [14] did not find any significant relationship between IT adoption and billing. They used the DRG coding ratio as a payment measure based on patient level and found that payments per discharge were essentially the same for IT adopters and non-adopters. They concluded that the hospitals considered in their studies were not systematically using EHRs to increase reimbursement.
These previous studies used specific functions, namely, EMR and CPOE, as measures of IT, although there may be more than 50 health IT systems in a hospital [16]. Thus, it would be somewhat difficult to separate specific functions from others, which could result in unobserved variable problems. Therefore, in this study, health IT was measured as a continuous variable, not just as a dummy variable as in previous studies [12,13,14]. Using a variable different from those used in previous studies, this study examined the effect of IT investment on the CMI by analyzing data obtained from the California Office of Statewide Health Planning and Development (OSHPD) from 2006 to 2007.

II. Methods

1. Data

We used hospital financial data obtained from the California OSHPD, as well as data from an American Hospital Association (AHA) survey and Healthcare Information and Management Systems Society (HIMSS) data. The California hospital financial data include patient utilization, hospital characteristics, and financial information. This data has been used in healthcare and economic studies [17,18]. The AHA data provide detailed statewide hospital information, such as hospital staffing, profiles, and utilization. The HIMSS provides a variety of detailed historical data, reports, and white papers about IT use in hospitals and integrated healthcare delivery networks on more than 5,000 US hospitals and the ambulatory facilities associated with these hospitals. In this study, data from 200 hospitals were considered from a 2-year period from 2006 to 2007.

2. Dependent Variable

The dependent variable was the CMI, which is a relative value assigned to the DRG of patients in a medical care environment. It is applied to determine the resource allocation to take care of the patients in a group [18]. For the CMI, each patient treatment record is assigned a Medicare Severity-DRG (MS-DRG) based on the patient characteristics. An MS-DRG has a weight representing the national average hospital resource consumption per patient for that group, relative to that of all patients [19]. The CMI was used as a payment measure to examine the relationship between IT adoption and billing [12,13,14].

3. Independent Variable

As the key explanatory variable, health IT expenditures were measured in dollar amounts, which were extracted from each hospital's trial balance worksheets and supplemental information sheets. IT expenditures include IT capital-related cost (i.e., physical capital, purchased services, lease/rental, and other direct expenditures) and IT labor-related cost (i.e., salaries and wages, employee benefits, and professional fees) [20,21].
The OSHPD data did not provide the adoption status of specific IT systems, such as EMR, CPOE, etc. To assess the validity of our IT capital measure, we examined its relationship to the discrete measures of more than 50 health IT systems (i.e. EMR, CPOE, PACS, patient billing, order entry, radiology information management, clinical documentation, etc.) based on information provided by the HIMSS, and we found that the adoption of IT systems is associated with IT capital investment. Thus, this measure of IT expenditure included all of the IT systems mentioned above, not specific ones.
We controlled two groups of independent variables, including hospital/market characteristics and volume of hospital service. The hospital/market characteristics included ownership (for-profit, not-for-profit, and public), teaching status (being a Council of Teaching Hospitals and Health Systems member), network hospital status (system member), number of beds (five specialized bed types: adult general acute, pediatrics, obstetrics, cardiac intensive, and neonatal intensive) and competition, which was measured according to the Herfindahl-Hirschman Index (HHI) for each hospital based on admissions given the geographical market of the health service area [17,18,20,21,22,23]. The values used to calculate volume included total admissions, outpatient visits, percentage of Medicare and Medicaid admissions out of total admissions, emergency room (ER) visits, and the numbers of inpatient and outpatient surgeries.

4. Statistical Analyses

To examine the effect of IT on CMI, this study utilized a generalized estimation equation (GEE) with log link and normal distribution. This estimation approach has been used in many studies focused on population-averaged effects [24]. For the covariance matrix, the independent variance model was adopted based on the smallest independence model criterion (QIC) [25]. The two groups of independent variables (hospital/market characteristics and volume) as well as IT expenditure were controlled. All analyses were conducted using Stata 11.2 software (StataCorp LLC., College Station, TX, USA).

III. Results

Table 1 shows the descriptive statistics for the variables used. The first row shows the CMI and IT investment. The average CMI was 1.106, but the standard deviation was a little low (0.256). Average IT investment was over 11 million. The second row of Table 1 shows the hospital characteristics. Not-for-profit hospitals accounted for almost 55% of the hospitals included in this study, while profit and public hospitals together accounted for 45%. Teaching and network hospitals accounted for 7.0% and 18.5%, respectively. Competition measured according to the HHI was 64.4%. The average number of specialized beds was 103 for adult general acute, 7 for pediatric general acute, 16 for obstetric, 5 for cardiac intensive, and 8 for neonatal intensive care. Network hospitals accounted for 18.8% of the hospitals considered.
The last row of Table 1 shows hospital volume. There were 10,370 total admissions and 147,375 outpatient visits. The percentage of Medicare admissions out of total admissions was 44.1%, and the percentage of Medicaid admission out of total admissions was 24.8%. The total number of ER visits was 31,902. The numbers of surgery inpatient and outpatient operations were 2,976 and 3,708, respectively.
Table 2 shows the variations of the key variables such as CMI and IT cost between 2006 and 2007. The average CMI increased by 1.5% from 1.098 to 1.114, and the average IT cost increased by 14% from $10,241,705 to $11,812,718.
Table 3 shows the GEE regression results. We found that IT was positively associated with CMI. For example, the CMI increased by 0.86% when IT investment increased by 10%. Hospital characteristics were also important factors in explaining CMI. Not-for-profit and public hospitals had lower CMI values than for-profit hospitals. However, teaching hospitals had higher CMI values than non-teaching hospitals. Only neonatal intensive beds were positively associated with CMI, while obstetric beds were negatively associated with CMI. Moreover, higher competition was associated with higher CMI values. In general, volume had a positive effect on CMI. The percentage of Medicare admission out of total admissions, ER visits, and inpatient surgeries were positively associated with CMI. However, the percentage of Medicaid admission out of total admissions was negatively associated with CMI.

IV. Discussion

This study examined the effect of health IT on CMI by considering hospital, market, and volume characteristics using data obtained from California hospitals during a two-year period. Unlike previous studies, we used continuous measurement of IT investment and found that IT was positively associated with CMI, indicating that increased IT adoption could lead to higher CMI values or higher billing though DRG up-coding.
Generally, a higher CMI in a hospital indicates that the hospital provides expensive services with higher coding and therefore receives more money from patients. Our results imply that hospitals could increase around $40,000 by increasing their IT investment by 10%, such an increase would represent a significant increase in profits for hospitals [26]. Our results are consistent with the findings of previous studies [12,13], which also found that IT adoption led to inflated reimbursement through DRG up-coding. These results confirmed that patient coding can be easily manipulated by using IT systems.
Regarding ownership, for-profit hospitals had higher CMI values associated with IT in comparison to not-for-profit and public hospitals. This finding is also consistent with those of other studies. Generally, for-profit hospitals are keen to earn profits, which may result from a higher CMI [12]. Teaching hospitals had higher CMI values because teaching hospitals may serve as referral centers for patients with severe diseases [27].
Among the types of beds, only neonatal intensive beds were positively associated with CMI. A neonatal unit provides mechanical ventilation, neonatal surgery, and special care for the sickest infants born in a hospital or transferred from another institution [28]. Thus, more neonatal intensive beds may result in a higher CMI. Network hospitals had higher CMI values. These hospitals may have had tougher cases or better medical record systems [29].
Among the volume variables, the percentages of Medicare, ER visits, and surgery inpatients are positively associated with CMI. Medicare patients are those who are older than 65 and may have more severe or chronic diseases. ER visits involve emergency situations and may require more resources than regular visits. Inpatient surgeries also require more resource to treat patients.
The results of this study suggest that the inflation of billing codes through health IT systems has a major impact on the healthcare industry. First, inflated billing leads to medical payment increases, which may not be justified by clinical benefits. Moreover, it may undermine risk-adjusted quality measurement. Finally, up-coding could degrade data integrity and the quality of care.
This study has important policy implications. Policy makers, researchers, and health professionals should be cautious in interpreting the effects of health IT on CMI, and they should remember that IT adoption itself could lead to higher patient billing. Thus, measures to prevent up-coding through IT systems should be implemented. For example, EMR systems should be monitored and audited at the payer side, and the way vendors design their EMR products should be regulated.


Conflict of Interest: No potential conflict of interest relevant to this article was reported.


1. Murphy EV. Clinical decision support: effectiveness in improving quality processes and clinical outcomes and factors that may influence success. Yale J Biol Med 2014;87(2):187-197. PMID: 24910564.
pmid pmc
2. Institute of Medicine, Committee on Quality of Health Care in America. Crossing the quality chasm a new health system for the 21st century. Washington (DC): National Academies Press; 2001.

3. El-Kareh R, Hasan O, Schiff GD. Use of health information technology to reduce diagnostic errors. BMJ Qual Saf 2013;22(Suppl 2):ii40-ii51.
crossref pmid pmc
4. Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005;293(10):1223-1238. PMID: 15755945.
5. Jones SS, Rudin RS, Perry T, Shekelle PG. Health information technology: an updated systematic review with a focus on meaningful use. Ann Intern Med 2014;160(1):48-54. PMID: 24573664.
crossref pmid
6. Agha L. The effects of health information technology on the costs and quality of medical care. J Health Econ 2014;34:19-30. PMID: 24463141.
crossref pmid pmc
7. Kuperman GJ, Gibson RF. Computer physician order entry: benefits, costs, and issues. Ann Intern Med 2003;139(1):31-39. PMID: 12834316.
8. Lainer M, Mann E, Sönnichsen A. Information technology interventions to improve medication safety in primary care: a systematic review. Int J Qual Health Care 2013;25(5):590-598. PMID: 23771745.
crossref pmid
9. Yu FB, Menachemi N, Berner ES, Allison JJ, Weissman NW, Houston TK. Full implementation of computerized physician order entry and medication-related quality outcomes: a study of 3364 hospitals. Am J Med Qual 2009;24(4):278-286. PMID: 19502568.
10. Himmelstein DU, Wright A, Woolhandler S. Hospital computing and the costs and quality of care: a national study. Am J Med 2010;123(1):40-46. PMID: 19939343.
crossref pmid
11. Lee J, Dowd B. Effect of health information technology expenditure on patient level cost. Healthc Inform Res 2013;19(3):215-221. PMID: 24175120.
crossref pmid pmc
12. Li B. Cracking the codes: do electronic medical records facilitate hospital revenue enhancement. Evanston (IL): Kellogg School of Management, Northwestern University; 2014.

13. Ganju KK, Atasoy H, Pavlou PA. Do electronic medical record systems inflate Medicare reimbursements?. Philadelphia (PA): Fox school of Business, Temple University; 2016. Fox School of Business Research Paper No. 16-008.

14. Adler-Milstein J, Jha AK. No evidence found that hospitals are using new electronic health records to increase Medicare reimbursements. Health Aff (Millwood) 2014;33(7):1271-1277. PMID: 25006156.
crossref pmid
15. Schulte F, Eaton J, Donald D. Doctors, others billing Medicare at higher rates [Internet]. Washington (DC): The Washington Post; 2012. cited at 2017 Oct 1. Available from:

16. Health Information Management Systems Society (HIMSS) Analytics [Internet]. Burlington (VT): HIMSS Analytics; 2016. cited at 2017 Oct 1. Available from:

17. Gowrisankaran G, Town RJ. Estimating the quality of care in hospitals using instrumental variables. J Health Econ 1999;18(6):747-767. PMID: 10847933.
crossref pmid
18. Lave JR, Pashos CL, Anderson GF, Brailer D, Bubolz T, Conrad D, et al. Costing medical care: using Medicare administrative data. Med Care 1994;32(7 Suppl):JS77-JS89. PMID: 8028415.
crossref pmid
19. Office of Statewide Health Planning and Development. Case Mix Index [Internet]. Sacramento (CA): Office of Statewide Health Planning and Development; 2016. cited at 2017 Oct 1. Available from:

20. Lee J. Impact of hospitalist care on hospital malpractice premiums using California hospital data. Appl Econ Lett 2017;24(11):742-752.
21. Lee J, McCullough JS, Town RJ. The impact of health information technology on hospital productivity. RAND J Econ 2013;44(3):545-568.
22. McCullough JS. The adoption of hospital information systems. Health Econ 2008;17(5):649-664. PMID: 18050147.
crossref pmid
23. Hayes KJ, Pettengill J, Stensland J. Getting the price right: Medicare payment rates for cardiovascular services. Health Aff (Millwood) 2007;26(1):124-136. PMID: 17211021.
crossref pmid
24. Gibbons RD, Hedeker D, DuToit S. Advances in analysis of longitudinal data. Annu Rev Clin Psychol 2010;6:79-107. PMID: 20192796.
crossref pmid pmc
25. Cui J. QIC program and model selection in GEE analyses. Stata J 2007;7(2):209-220.
26. Association of Clinical Documentation Improvement Specialists. What does case-mix index mean to you? [Internet]. Middleton (MA): HCPro; 2010. cited at 2017 Oct 1. Available from:

27. Mendez CM, Harrington DW, Christenson P, Spellberg B. Impact of hospital variables on case mix index as a marker of disease severity. Popul Health Manag 2014;17(1):28-34. PMID: 23965045.
crossref pmid pmc
28. American Hospital Association. AHA hospital statistics. Chicago (IL): Health Forum LLC; 2017.

29. Institute of Medicine, Committee on Quality of Health Care in America. For-profit enterprise in health care. Washington (DC): National Academies Press; 1986.

Table 1

Descriptive statistics for California hospitals from 2006 to 2007


Values are presented as mean ± standard deviation.

NFP: non-for-profit, ER: emergency room.

Table 2

Case Mix Index and IT cost variation between 2006 and 2007


Values are presented as mean ± standard deviation.

Table 3

GEE regression results with log link and normal distribution for California hospitals from 2006 to 2007


GEE: generalized estimation equation, NFP: non-for-profit, ER: emergency room, SD: standard deviation.

*p<0.1, **p<0.05, ***p<0.01.


Browse all articles >

Editorial Office
1618 Kyungheegung Achim Bldg 3, 34, Sajik-ro 8-gil, Jongno-gu, Seoul 03174, Korea
Tel: +82-2-733-7637, +82-2-734-7637    E-mail:                

Copyright © 2020 by Korean Society of Medical Informatics. All rights reserved.

Developed in M2community

Close layer
prev next