In this cross-sectional survey performed in a primary care setting, QoL in patients with type 2 diabetes is significantly lower compared to the general population. Additionally, this study revealed declining scores for all SF-36 subscales with an increasing number of comorbid conditions. The most common comorbid conditions reported were hypertension and osteoarthritis with osteoarthritis having remarkable more impact on quality of life than hypertension.
Over the last two decades health related quality of life, individual health status or well-being have gained more importance as patient-relevant outcome parameters within medical and health services research . Especially for patients suffering from one or several chronic conditions care should focus on the best possible management of the disease and additional impairments on daily life instead of recovery and health. [2, 20]. For older patients improvements within QoL may often have a more important role than a possible extension of life time ("add life to years, not years to life") [21, 22].
Comparable to results of other studies [3, 23–25] patients with type 2 diabetes in our sample were limited in all scores of the SF-36 compared to people without diabetes. According to the literature the number of comorbid conditions was associated with a lower quality of life in all domains of the SF-36 [26, 27]. Interestingly in our study patients with hypertension and diabetes achieved higher scores than patients with only diabetes. However, these differences did not reach statistical significance after adjusting for relevant variables. These findings are in accordance with previous studies, describing similar quality of life scales of patients with hypertension and those without any chronic condition [28, 29]. One reason for this finding may be that hypertension is often asymptomatic and physically less impairing than other diseases. However, other studies showed hypertensive patients to have lower scales in QoL than normotensive patients because of adverse effects of drugs used in the treatment of the high blood pressure  or because of a so called labelling effect . Wee et al. assumed that there are chronic conditions with non-additional effects on health related QoL, so that having both conditions is not more disabling than having one of them . Sprangers et al. describe a mechanism of accommodation to a chronic illness with changes in internal standards and values – the so called "response shift" .
It is important to keep in mind that hypertension perhaps does not intensify the burden for the patients since high blood pressure levels represent a major risk factor for cardiovascular mortality and morbidity especially for patients with type 2 diabetes . This has to be taken into account as an additional and important risk factor, both from patients and from physicians .
Regarding osteoarthritis as comorbidity we found remarkable lower scales in all domains of the SF-36 in particular within the subscales related to physical well-being. The revealed high burden of patients with osteoarthritis is in accordance with other studies and congruent with the clinical experience of primary care physicians [33–36]. Major problems for patients with osteoarthritis are pain and disability. These symptoms are associated with an increased health service utilization [35, 37, 38] and have to be kept in mind when dealing with diabetic patients with concomitant osteoarthritis.
The list of self reported comorbidities used in this survey did not contain any mental conditions like e.g. depression, so we were not able to assess the possible impact of these potential comorbidities as we did with somatic comorbidities. However, the used set of questionnaires contained the PHQ-9 as a screening instrument for depressive disorder. This enabled us to control our data for this important issue [12, 26]. To evaluate the impact of mental comorbidity on QoL in primary care further research is still needed.
The present study has some limitations. First of all the results were cross-sectional, any conclusions on causality are impossible. All data were self reported, some chronic conditions could be under- or overreported. All questions were filled out self-dependent, considering the mean age of the participants misconceptions could not be excluded. Furthermore calculating the BMI out of self reported height and weight is associated with a limited validity especially in older adults [39, 40]. Smoking rates in our sample were self reported too. But there is some evidence that the validity of self-reported smoking within survey studies is reasonable . Furthermore the BMI and the percentage of smokers in our study sample were comparable to findings in the primary care population in the US and Germany [42–44].
The most important limitation might be that we had no knowledge about the severity of the addressed comorbidities. A fact which might limit generalizability of our findings is that all participants of our survey were from the same regional health fund. This insurance fund covers a sample with a higher proportion of elder insurants and a higher prevalence of multimorbidity than other insurers in Germany.
The response rate of our survey was moderate, but a non-responder analysis could be performed, showing that non-responder were slightly older and more likely to be female. The response rates might have been higher if the questionnaires would have been sent out by the university department directly  instead of the health insurance fund. However, due to a strict protection of data privacy we weren't able to contact the patients directly.
Strengths of our study were the large and heterogeneous study sample collected in a primary care setting. Since patients' selection was primarily conducted by using routine claims data and secondarily by drawing a random sample selection bias is unlikely.