Structured treatment programs in morbidity-oriented risk structure compensation using the example of diabetes and COPD

As of January 1, 2009, the risk structure equalization (RSA) in statutory health insurance was further developed into a morbidity-oriented RSA for 80 serious or cost-intensive chronic diseases (Göpffarth 2009). Since 2003, however, there has been a preliminary stage of morbidity orientation in the form of the coupling of structured treatment programs (also disease management programs - DMP) to the RSA (Busse 2004, Stock et al. 2006). The background was a discussion about the control effect of the (old) risk structure compensation. Among other things, it was criticized that health insurance companies have no (financial) incentives to implement structured treatment programs under the given conditions. If they implemented appropriate programs, the argument goes, they would become attractive to chronically ill people, who would then increasingly switch to health insurance. However, since chronic illnesses are not taken into account in the RSA, these switches would result in negative contribution margins and would thus lead to the financial ruin of the health insurance company.

>> In 2001, the legislature responded to this by enacting a legal roadmap for the introduction of morbidity-based risk adjustment. At the same time, it introduced Disease Management Programs (DMPs) as an interim solution. These programs address selected chronic diseases (bronchial asthma, breast cancer, COPD, type 1 and type 2 diabetes, and coronary heart disease) and are subject to legally binding requirements. The programs must be approved by the Federal Insurance Office (BVA) before insured individuals can enroll. Within the risk adjustment scheme, separate groups of enrolled insured individuals were then created based on their specific disease, and a separate contribution requirement was calculated for each group. In addition, a flat-rate DMP program fee of approximately €86 per enrolled insured individual was introduced to compensate for the increased administrative costs of implementing the programs.
These were, overall, very effective financial incentives for implementing disease management programs (DMPs). After the first programs were launched in 2003, a total of 2.2 million insured individuals had enrolled by the end of 2005. Enrollment increased steadily in subsequent years: 3.1 million (end of 2006), 4.3 million (end of 2007), and 5.5 million insured individuals (end of 2008).
With the introduction of the morbidity-based risk structure compensation scheme (RSA), the separate consideration of disease management programs (DMPs) is no longer required. The new morbidity surcharges, however, are calculated for all individuals affected by a disease, regardless of whether they are enrolled in a program or not. A modified program cost allowance remains, amounting to €180 per enrolled insured person (from 2011 onwards: €168).
This has significantly altered the financial incentives for health insurance funds to implement disease management programs (DMPs). Many feared that the programs would no longer be continued under the changed conditions—or only in a modified form (Stock et al. 2006, Straub/Dietrich 2008, Lüdtke-Handjery 2008). However, this did not hinder the growth in enrollment figures. During 2009, enrollment continued to rise to 6.2 million insured individuals, and by the end of 2010 to 6.5 million. This article presents the different considerations of DMPs and the underlying indications in the previous and new risk adjustment schemes and compares the form of consideration using the examples of diabetes mellitus and COPD. Finally, the main findings are summarized.
The figures presented here are based on the 2008 risk adjustment annual settlement, which was still carried out under the old regulations, and are compared to a hypothetical 2008 annual settlement according to the morbidity-based risk adjustment scheme (legal status 2010). For better comparability, the values of the 2008 risk adjustment annual settlement are presented without taking the risk pool into account.
Consideration of the DMP in the RSA
The socio-demographic RSA 2003-2008
The old risk adjustment scheme (RSA) was based on sociodemographic characteristics such as the insured person's age and gender, as well as their status as a recipient of a reduced earning capacity pension (EM status). The calculation was performed using a cell-based approach, meaning that the insured person was assigned to a unique RSA cell based on their characteristics, allowing the contribution requirement to be determined as the average benefit expenditure per RSA cell.
In 2003, enrollment in a Disease Management Program (DMP) was added as a further morbidity marker. Separate insured groups were formed for each program, segmented by age, gender, and disability status. The number of risk adjustment (RSA) cells thus increased rapidly, from 670 cells to a final 4,355 cells (Otto/Göpffarth 2010: 175). However, the assignment to an RSA cell must be unambiguous. Specifically, this means that an insured person with multimorbidity could enroll in more than one DMP (e.g., COPD and diabetes), but could only be considered for RSA purposes for one program.
The financial impact of Disease Management Programs (DMPs) under the old risk adjustment scheme (RSA) was often overestimated. For example, the €5,521 annual payment for an enrolled insured person with COPD in 2008 should not be compared to the €1,811 received by a non-enrolled insured person. Rather, insured persons with COPD are older than the average statutory health insurance (SHI) member and would have already received higher premiums due to their age, even without DMP affiliation. The correct comparison is therefore the amount an age-matched group would receive, namely €3,482. The increased premium for enrolled insured persons was financed by reducing the premium for non-enrolled insured persons from €1,906 to €1,811. However, this is only the average reduction, the specific amount of which depended on the particular age and gender group. Whether a health insurance fund benefited from linking the DMP to the RSA therefore depended less on the raw enrollment rate, but on the age- and gender-adjusted enrollment rate (Otto/Göpffarth 2006).
Since the reduction in contribution requirements for non-enrolled insured persons affected all health insurance funds – regardless of DMP enrollment – it was indeed important for the health insurance funds to realize as high a proportion of their enrollment potential as possible – critics objected to this as an enrollment competition (Straub 2005), while proponents welcomed it as a broad approach (Stock et al. 2007).

The morbidity-oriented RSA since 2009
In 2009, the risk structure compensation (RSA) system was further developed into a morbidity-based RSA. In addition to the previous criteria of age, gender, and disability status, morbidity groups were added for 80 selected serious and costly chronic diseases. All indications for which a disease management program (DMP) existed before 2009 are among the 80 selected diseases (Göpffarth 2008). Assignment to the insured group is based on the diagnoses documented by physicians and reported to the health insurance funds as part of the billing process. While hospital diagnoses lead directly to the surcharge, further criteria apply to diagnoses from outpatient care. For example, the diagnoses must generally have prompted treatment in two different billing quarters. For some diseases, a comparison is also made with the prescribed medications, e.g., insulin for type 1 diabetes mellitus.
The allocations that a health insurance fund receives from the health fund for an insured person consist of a base amount differentiated by age and gender, as well as, where applicable, surcharges for disability status and/or assignment to morbidity groups. The basis for assignment is therefore no longer enrollment in a program, but rather the existence of a corresponding diagnosis. While the separate disease management program (DMP) contribution requirements were differentiated by age and gender, this is only the case for morbidity surcharges in exceptional cases. However, the morbidity surcharges for a disease are sometimes differentiated by severity. For diabetes, there were four different surcharges differentiated by severity in 2010. In these cases, the surcharges are arranged hierarchically, so that a surcharge is only granted for the most severe documented manifestation of the disease. In total, 112 different surcharges were differentiated for the 80 diseases in 2010.
The calculation of the surcharges is now based on a regression method rather than a cell-based approach. This allows for multiple assignments of insured individuals, enabling the combination of surcharges in cases of multimorbidity.
This assumes that expenditures in multimorbidity are linearly additive, which should first be critically examined. Empirically, however, non-linear effects appear to be relevant only in a few cases (Schulte, 2010).
3. Comparison of the procedures using the examples of diabetes and COPD
Mapping of diabetes and COPD in the morbidity-oriented RSA
The way diabetes and COPD are considered in the classification model (as of 2010) can be seen in Figure 1. Accordingly, for diabetes, depending on the severity, there is a morbidity supplement of between €564 and €2,202, which is paid in addition to the basic contribution based on age and sex. For COPD, the grouping clearly shows that economic cost homogeneity is an important criterion in the formation of the morbidity groups. COPD is found here, sometimes together with bronchial asthma, status asthmaticus, and emphysema, in HMG 109 and 110. A separate analysis will therefore not be possible in the following. Asthma/COPD is one of the few cases in which an age split was performed (Fig. 1).
For uncomplicated diabetes, no intervention criteria have been defined; that is, two diagnoses from outpatient care or a hospital diagnosis must be present. If complications are present, however, medication validation with insulin or oral antidiabetic drugs (> 182 DDD) is performed. In the case of type 1 diabetes mellitus, insulin must have been prescribed. In the case of diabetic ketoacidosis or coma, a corresponding hospital diagnosis must be present. No intervention criteria have been defined for cystic fibrosis, emphysema, and chronic obstructive bronchitis. However, in the case of bronchial asthma, medication with at least 183 daily doses must be present, and in the case of status asthmaticus, a corresponding hospital diagnosis is required.
Comparison of contribution requirements and surcharges
To facilitate comparison with the old risk adjustment scheme (RSA), the figures are presented as an example for a 62-year-old woman. Under the old RSA, a health insurance fund received a contribution requirement of €3,599 for a 62-year-old woman enrolled in the Disease Management Program (DMP) for type 2 diabetes. In the morbidity-based RSA, there is a base amount of €1,210. Together with the diabetes supplement, the allocation ranges from €1,774 (without complications) to €3,412 (with renal or multiple complications). Only in cases of the most severe complications are the values comparable to those in the old RSA. For COPD, the contribution requirement for a 62-year-old woman is €4,220. In contrast, the allocation in the morbidity-based RSA is €2,093 (€1,210 + €883).
The difference arises from the fact that, in the morbidity-based risk adjustment scheme (RSA), in addition to the supplement for the primary disease, supplements are also granted for other documented diseases. Figure 2 shows the average allocations (average across all age groups and severity levels) for the individual morbidity groups. Overall, the average allocations for type 2 diabetics (HMG 015 to 019) range from €4,478 to €10,104, for type 1 diabetics (HMG 020) €8,333, and for asthma/COPD (HMG 109/110) from €1,543 to €4,777. The comparable average values for the disease management programs (DMPs) are €5,075 (type 2 diabetes), €6,172 (type 1 diabetes), €5,521 (COPD), and €2,889 (asthma).
The average values considered should not obscure the variance in individual cases. While only minor differences based on age and gender were taken into account in the Disease Management Programs (DMPs), the range is considerably larger in the morbidity-oriented Risk Adjustment Scheme (RSA). For example, if an insured person with uncomplicated diabetes (HMG 019) has no other illnesses, the allocation is reduced to approximately half (2,206 euros). The most important comorbidities for diabetes and COPD, as well as their frequencies, can be found in Table 1. For instance, more than 45% of the individuals assigned to a diabetes morbidity group also receive a supplement for hypertension (Table 1).
Comparison of the affected groups of people
In addition to comparing the allocations of the morbidity-based risk adjustment scheme (RSA) with the contribution requirements of the old RSA, a comparison of the affected groups of people is also of interest. While the old RSA relied on the insured person's enrollment in a disease management program (DMP), the documented diagnoses are now the determining factor. This should lead to an expansion of the group of people considered. On the other hand, the question arises whether there are enrolled insured persons who do not receive a morbidity supplement. It is also of interest whether the enrolled group of people differs from the unenrolled group.
Tables 2 and 3 provide corresponding cross-tabulations for type 2 diabetes mellitus and asthma/COPD. The total number refers to the sample dataset for the risk adjustment scheme (RSA), which comprises 4.3 million insured individuals. Of these, nearly 330,000 insured individuals, i.e., approximately 7.6 million, either receive a morbidity supplement for type 2 diabetes mellitus or are enrolled in the corresponding disease management program (DMP). Slightly more than half of all diabetics are enrolled in the DMP. Nearly 156,000 diabetics who are not enrolled in a DMP receive a morbidity supplement for diabetes. However, 8.9 million of these diabetics do not receive a morbidity supplement. This is presumably because the disease was not documented in two billing quarters (for the underreporting of diabetes, see Erler et al., 2009).
A similar picture emerges for asthma/COPD. Here, the prevalence is 4.9 per 1,000,000 people. However, only about 20 per 1,000,000 people with this condition are enrolled in the corresponding Disease Management Programs (DMPs). It is striking that almost a third of those enrolled in DMPs do not receive the corresponding morbidity supplement. This is because medication validation is performed for asthma. Asthmatics without medication can enroll in the DMP, but are not considered in the morbidity-based risk adjustment scheme.
The question of whether patients enrolled in a Disease Management Program (DMP) differ from those not enrolled is of interest. Häußler et al. (2005) conclude that "younger patients and those with less severe illnesses are overrepresented in the programs. Older patients and those with more severe illnesses are thus under-reached by the programs." To test this hypothesis, the study examines whether the allocations for enrolled DMP patients differ systematically. While type 2 diabetics receive an average allocation of €5,444 overall, the figure for enrolled diabetics is €5,207. This initially supports the hypothesis.
However, a further breakdown reveals a slightly different picture. The AGG and EMG allocations are marginally higher for enrolled insured individuals (1,588 euros compared to 1,580 euros and 81 euros compared to 80 euros, respectively). With regard to age and gender distribution and receipt of disability pensions, there is no difference between enrolled and non-enrolled insured individuals. The average morbidity supplement for diabetes is also higher for enrolled insured individuals (914 euros compared to 902 euros). Enrolled diabetics thus tend to have a slightly higher (documented) severity level of diabetes. However, the average allocations for other illnesses for enrolled diabetics are lower than those for all diabetics (2,623 euros compared to 2,882 euros). The key difference, therefore, lies in the fact that enrolled diabetics exhibit below-average multimorbidity.
Even in cases of asthma and COPD, the average allocations for all affected individuals are considerably higher at €4,797 than for enrolled insured individuals at €3,577. However, this result is likely skewed by the fact that drug validation is carried out for asthma, meaning that milder cases can enroll but do not receive a supplement.
Conclusions
The financial framework for structured treatment programs has changed considerably with the introduction of morbidity-based risk adjustment. While all disease management program (DMP)-relevant indications are considered for the selected diseases, the way they are considered differs significantly in some cases. Key differences include the criteria for inclusion, the differentiation of severity levels, and the consideration of additional diseases.
The specific morbidity surcharges cannot be compared with the old DMP contribution requirements; rather, the total amount of allocations must be considered, broken down by age and gender, disability status, and including morbidity surcharges for other illnesses. This achieves a much greater differentiation in the allocation amount, better addressing varying degrees of severity and comorbidities. Particularly in cases of multimorbidity, this results in significantly higher targeting accuracy of allocations, even though the consideration of multimorbidity in the morbidity-oriented risk adjustment scheme still has room for improvement (Schulte 2010).
In particular, the morbidity surcharges cover broader patient groups than the disease management program (DMP) contribution requirement, which was limited to enrolled insured individuals. However, from the DMP's perspective, this also poses a risk, as it eliminates the special financial incentive for implementing the programs. On the positive side, this means that genuine competition between different forms of care can now take place. On the negative side, it means that allocation no longer depends on the quality of care. Whether the "power of demand" of those affected in this competition will be sufficient as a corrective – since, from the health insurance company's perspective, they no longer represent high-risk patients – remains to be seen. The continued flat-rate DMP program cost payment can be seen as an initial – and expandable – element of quality improvement. Given the continued increase in enrollment numbers, this payment appears to be sufficient to keep the programs attractive to health insurance companies.
The striking finding is that while those enrolled in a Disease Management Program (DMP) are not younger or less seriously ill than those not enrolled, they are less likely to have multiple comorbidities. This aligns with previous financial incentives that did not take multimorbidity into account. This may represent the greatest and most important potential for further developing DMPs.