Have DRG-based prospective payment systems influenced the number of secondary diagnoses in health care administrative data?
Introduction
In Sweden, the number of secondary diagnoses per case in inpatient care increased during the 1990s. Reasons for this could include hospitalisation of patients with more severe diseases. Another assumption is that the introduction of DRG-based prospective payment systems in hospital care has contributed to an increase in recording diagnoses, and consequently increased the number of secondary diagnoses per case. One part of the health care reform in the 1990s in Sweden was to introduce per-case payment in the new health care purchaser-and-provider model [1]. Stockholm was the first county council to adopt the new concept of using prospective payment based on diagnosis-related group (DRG) systems [2].
DRGs are secondary patient classification systems based on primary classified medical data grouping single events of care into larger, economically and medically consistent groups with a previously defined algorithm. The main primary classified medical data are diagnoses and surgery codes.
Apart from their use in reimbursement systems, case-mix systems such as DRG were designed for planning, budgeting, management and financing inpatient care.
Introduction and application of a DRG system focuses on recording diagnoses. As DRG systems are adopted as reimbursement systems, hospitals evidently become more interested in the coding processes—the economic results come more or less to depend on coding practice.
Few studies—and none outside Sweden—describe the changes in diagnosis frequency when DRG-based prospective payment systems have been introduced. A study by Paulson [4] describes the consumption of hospital treatment in relation to the introduction of the Stockholm model in the County Council of Stockholm. Svensson and Garelius [5] studied the effect of economic incentives on physicians' decisions in the Stockholm County Council area. They found that physicians got incentives to code secondary diagnoses more carefully—if the relevant secondary diagnoses were not coded, the hospital department got less money. In both studies, the number of secondary diagnoses increased after prospective payment was introduced. Other studies, presented in international journals, emphasise hospital care quality after the introduction of DRG-based prospective payment systems ([6], [7], [8]). Studies in Sweden with the same purpose have been presented by Forsberg [9] in studies of the Stockholm model.
However, there may also be other reasons for the increase in more diagnoses. Greater interest in follow-up of health care after the introduction of DRG as a management system may have stimulated the recording of diagnoses. A gradual change in the mix of specialities could have increased the number of secondary diagnoses. There may have been case selection: movement of less sick patients from large, highly specialised, regional hospitals to smaller county hospitals may have increased the number of secondary diagnoses at the regional hospitals. Hospital patients may have got sicker, increasing diagnosis frequencies in general. One possibility is that economic incentives from prospective systems increase the recording of diagnoses. Another question is whether the increased recording of diagnoses depends on improvement in recording or an abuse of coding practice. Additionally, the change from the ICD-9 coding system to ICD-10 [3] in 1997 could also have increased the interest in recording diagnoses. Some of the above plausible reasons can be investigated; others are more difficult to evaluate.
The principal objective of the present study was to investigate whether introduction of DRG-based prospective payment systems in Sweden has influenced the frequencies of secondary diagnoses in administrative data.
Section snippets
Method
The nation-wide Hospital Discharge Register between 1988 and 2000 was used for analysis. This register covers all inpatient care in Sweden, both public and private. The register includes patient particulars such as age, sex, the unique personal identification number, when and where the care took place and the reasons for hospital treatment (diagnoses, surgical procedures, etc.).
The year 1988 was chosen as a baseline date to eliminate the risk of influence from prospective payment systems at the
Results
Secondary diagnoses increased from 0.44 to 0.84 per case between 1988 and 2000, an increase of 94% (Fig. 1).
The number of secondary diagnoses per case can be considered: (a) as the share of cases with secondary diagnoses and (b) the numbers of secondary diagnoses for each case that had secondary diagnoses coded.
In 1988, 29% of the cases had secondary diagnoses coded; by 2000, the proportion had risen to 46%. For cases with secondary diagnoses, number of secondary diagnoses increased from 1.52
Discussion
The results show that hospitals with DRG-based prospective payment systems had a steeper increase in the number of secondary diagnoses per case starting from when their prospective payment system was introduced. Regional hospitals without prospect payment systems had a more constant increase; it came later and coincided with their introduction of DRG-based management systems. Other reasons may have contributed to the increase of secondary diagnoses, as discussed below.
Has the severity of
Conclusion
The recording of secondary diagnoses increased during the 1990s, more or less for all regional hospitals in Sweden. Introducing DRG-based prospective payment system just to increase the recording of diagnoses is not a goal in itself, but a DRG-based system gives the positive effect of more diagnoses per case, which hopefully gives a more comprehensive description of health care.
Four hypotheses were formulated:
- •
The introduction of DRG-based prospective payment systems increases secondary
References (14)
The monopolistic integrated model and health care reform: the Swedish experience
Health Policy
(1996)- NordDRG Users Manual. Nordic Centre for the Classification of Diseases and the national NordDRG owners, Version...
- Klassifikation av sjukdomar och hälsoproblem. Systematisk förteckning. (Swedish version of International Statistical...
- Paulson E. Stockholmsmodellen och fem kirurgiska kliniktyper-en longitudinell studie av vårdkonsumtion. Rapport...
- Svensson H, Garelius L. Har ekonomiska incitament påverkat läkarnas beslutsfattande? (Have economic incentives had an...
- et al.
The effects of the DRG-based prospective payment system of quality of care for hospitalized Medicare patients. An introduction to the series
The Journal of the American Medical Association
(1990) - et al.
Quality of care before and after implementation of the DRG-based prospective payment system. A summary of effects
The Journal of the American Medical Association
(1990)