This study showed that type 2 diabetes patients with a comorbidity had a lower health status than patients without. Furthermore, health status decreased with an increasing number of comorbidities, except for the mental health measures. Patients with both cardiovascular and non-cardiovascular comorbidities, had a significantly lower physical health than patients with only cardiovascular or non-cardiovascular comorbidity. The percentage of type 2 diabetes patients with one or more comorbidities in our study was 38%, while in another study in the Netherlands this was 44%. This difference is probably due to our selection of well-controlled, no insulin using diabetes patients.
While the link between diabetes patients with comorbid diseases and poor health status is well established, our study is novel because it looked at a group of people with well-controlled type 2 diabetes. Our results are consistent with findings from general diabetes populations. We found that comorbidity in diabetes patients was mainly associated with diminished physical health, as was described for osteoarthritis, stroke, cardiovascular disease, respiratory disease and myocardial infarction[3–7, 11]. Furthermore, we also found that a higher number of comorbidities is associated with a decreased health-related quality of life in type 2 diabetes patients[7, 8, 11, 13, 14].
In contrast to aforementioned types of comorbidities, comorbid depression is associated with both a lower physical and mental health status[5, 6, 8, 10, 12]. This means that the mental health status scores are likely to be associated with depression, however we did not measure depression. We found that a higher number of comorbidities did not further decrease mental health, role limitations due to emotional problems and mental component score. Furthermore, the Mental Component Scores in our population were rather high (mean: 54) indeed and patients with scores above 42 are not likely to have a depression. These findings suggest that depression (and thus a decreased mental health) was not a major issue in our study population. This is in accordance with the fact that depression is associated with poor glycemic control and we studied a selection of well-controlled diabetes patients.
The decrease in health status with increasing number of comorbidities seen in most SF-36 domains was also seen in the EQ-5D and EQ VAS. However, these changes were not significant. This could be explained by the fact that these health status measures combine both physical and mental health in one score. Therefore, the differences between one, two and more than two comorbidities might be diluted by the large differences for physical health on one side and almost no differences for mental health on the other. Furthermore, this could be due to the relatively small groups of people with two and more than two comorbidities. Also, we found rather high values for health status. This could be due to the fact that poor self-rated health is associated with increasing glucometabolic disturbances and such patients were not included in the study.
Also patients with diabetes and microvascular complications[7, 10, 34, 35] have a lower health-related quality of life compared to patients with diabetes alone; however we did not look at the association between microvascular complications and health status. Patients with end stage renal disease were not included, because they are not treated by their general practitioner. The same applies to patients with diabetic foot disease or neuropathy. Retinopathy is estimated to be present in about 7.4% of the type 2 diabetes patients in the Netherlands. Because our population had a relatively short diabetes duration and people were well-controlled, our retinopathy prevalence would have been low. This was confirmed by the fact that retinopathy was only mentioned nine times among ‘other diseases’. Therefore we think this will not have influenced our results.
Several studies studied the relation between comorbidity and glycemic control and since we looked at comorbidity in a selection of well-controlled type 2 diabetes patients, their findings might be of interest to interpret our findings. The risk of cardiovascular diseases increases with a higher HbA1c, so our study population had a relatively low risk of getting ‘new’ complications. On the other hand, diabetes patients with coronary heart disease or congestive heart failure have a lower odds of having at least one cardiovascular risk factor (glycemia, blood pressure and lipids) out of control. Since our results are similar to studies in general (both well-controlled and not well-controlled patients) diabetes populations, we think that ‘good cardiometabolic control’ was not a confounder in the association between comorbidity and health status. In the relationship between diabetes control and comorbidities, general practitioners’ treatment and patients’ health behavior are likely to play a role as well. Because of the homogeneity of the study population (all were well-controlled) we were not in the position to elucidate whether the health status of patients is associated with patient’s behavior.
Although quality of care may increase with an increasing number of chronic conditions, our study demonstrated that acceptable or good cardiometabolic control does not automatically reflect a good health status. So far, three types of interventions were designed to improve quality of life in diabetes patients: a disease management program, implementing several elements of the Chronic Care Model and a structured group self-management educational intervention. Diabetes patients in the German disease management program had higher health-related quality of life in the dimensions mobility, self-care and performing usual activities compared to routine care. The same was shown for patients with a high number of comorbidities. The disease management program also improved processes of care and intermediate outcomes. However, there were no differences in intermediate outcomes between the disease management program and routine care. The number of secondary diseases and the presence of a disabling secondary disease were related to drop-out of the disease management program. This disease management program showed that it may improve health-related quality of life and therefore may be useful in clinical practice. However, one might question their usefulness for diabetes patients with comorbidities. Disease management programs focus on a single disease and as we could demonstrate health status does not depend on a single disease. In our opinion the above mentioned drop-out is not surprising; in disease management programs special attention should be paid to patients with both cardiovascular and non-cardiovascular comorbidity. The other two mentioned interventions are still ongoing.
Strengths of this study are that we have measured multiple health status domains and multiple comorbidities. On top of that, we have measured them all in one, large population. This allowed us to quantify the impact of comorbidity on several health status domains. This could be helpful in providing specific treatment options to type 2 diabetes patients, depending on the different comorbidities and the different aspects of their health status. Because the included patients were not only recruited for a randomised controlled equivalence trial but were also part of a patient preference study, we minimised selection bias. Besides, we had a high response rate on the patient questionnaires, probably due to the fact that patients had already agreed to participate in a larger study and thus were more motivated to fill in the questionnaire.
There are also limitations that need to be addressed. Firstly, comorbidity in our study population might be underreported, and lack of disease coding in the electronic medical records might play a role. We cannot assess its role, but underreporting is not likely to influence the direction of our results. Secondly, we selected only eleven conditions, seven of which were vascular diseases. Conceptually, multimorbidity includes all potential other conditions. However, other diseases were hardly reported as meaningful diseases, with the exception of cancer. Therefore we assume that this will not have biased our results. Thirdly, we have no information about the medication used for the comorbidity. So we do not know if it is the comorbidity itself that decreased health status or the medication patients are using for it. Fourthly, we wanted to assess the impact of comorbidity regardless of possible confounders such as age, gender, ethnicity and educational level of patients. However, adjusting for these possible confounders implies that the impact of these confounders on health status could not be assessed separately. Lastly, because of the cross-sectional design no causal relationship could be established. However, a cohort study with a follow-up of five years showed that in elderly diabetes patients the diabetes-related complications were predictive of reduced health-related quality of life, but the number of comorbid diseases did not. We found that both diabetes-related complications (cardiovascular comorbidity) and the number of comorbidities were associated with a reduced health status, so the cohort study only partly confirms to our results. The fact that in the cohort study the number of comorbid diseases did not predict reduced health-related quality of life may be explained by the fact that at baseline health-related quality of life was already lower in diabetes patients with comorbid diseases, therefore during the study period no further reduction in health-related quality of life might be detectable.