[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.205.176.107. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Figure. Predictors of patient complexity adjusted for patient age, sex, and hemoglobin A1c (HbA1c) level. The adjusted odds ratios are derived from separated logistic regression models after adjustment for age, sex, mean HbA1c level in the previous year, and patient clustering by primary care physician. COPD indicates chronic obstructive pulmonary disease.

Figure. Predictors of patient complexity adjusted for patient age, sex, and hemoglobin A1c (HbA1c) level. The adjusted odds ratios are derived from separated logistic regression models after adjustment for age, sex, mean HbA1c level in the previous year, and patient clustering by primary care physician. COPD indicates chronic obstructive pulmonary disease.

1.
Jackson GL, Edelman D, Weinberger M. Simultaneous control of intermediate diabetes outcomes among Veterans Affairs primary care patients.  J Gen Intern Med. 2006;21(10):1050-1056PubMedArticle
2.
Safford MM, Allison JJ, Kiefe CI. Patient complexity: more than comorbidity: the vector model of complexity.  J Gen Intern Med. 2007;22:(suppl 3)  382-390PubMedArticle
3.
Grant RW, Ashburner JM, Hong CC, Chang Y, Barry MJ, Atlas SJ. Defining patient complexity from the primary care physician's perspective: a cohort study.  Ann Intern Med. 2011;155(12):797-804PubMed
4.
Piette JD, Kerr EA. The impact of comorbid chronic conditions on diabetes care.  Diabetes Care. 2006;29(3):725-731PubMedArticle
Research Letters
May 28, 2012

Characteristics of “Complex” Patients With Type 2 Diabetes Mellitus According to Their Primary Care Physicians

Author Affiliations

Author Affiliations: Division of Research, Kaiser Permanente Northern California, Oakland, California (Dr Grant); and Diabetes Center (Dr Wexler) and Division of General Medicine (Mr Ashburner and Drs Hong and Atlas), Massachusetts General Hospital, and Harvard Medical School (Drs Wexler, Hong, and Atlas), Boston, Massachusetts.

Arch Intern Med. 2012;172(10):821-823. doi:10.1001/archinternmed.2012.1229

Despite recent trends toward improved risk factor control, most patients with type 2 diabetes mellitus still do not achieve all evidence-based management goals,1 suggesting that new approaches are needed to further improve the quality of diabetes care. Patient complexity is a concept that is defined to describe the multiple factors that contribute to the challenges associated with clinical care.2 Because primary care physicians (PCPs) have a unique perspective on type 2 diabetes management, we hypothesized that greater insight into PCP-defined complexity among patients with type 2 diabetes could help inform strategies for improving diabetes primary care.

Methods

We conducted a cross-sectional analysis of PCP-defined patient complexity by asking 40 PCPs from the Massachusetts General Primary Care Practice-Based Research Network, Boston, to review randomly generated lists of 120 of their own patients and to designate which of these patients “in their view” they considered complex.3 Among the patients with type 2 diabetes, we examined the relative impact of diabetes-related vs other more general medical conditions on PCP-defined complexity. For each comorbid diagnosis that was significantly associated with complexity in univariate analysis, we constructed a separate logistic regression model and reported the relative odds of PCP-defined complexity after adjusting for age, sex, glycemic control, and patient clustering by PCP. The study was approved by the institutional review board of Massachusetts General Hospital, Boston.

Results

The PCPs reviewed 327 patients with type 2 diabetes (mean [SD] number of patients per PCP, 8.4 [5.1]; range, 2-25 patients per PCP) and designated 68.2% of these patients as complex. The PCP-defined complex patients with diabetes were 3 years older and more often female than noncomplex patients with diabetes but had similar race/ethnic prevalence and lived in neighborhoods with similar median household incomes. Complex patients with diabetes made more clinic visits and were prescribed a greater number of different medications (mean [SD] number of medications, 20.1 [10.0] vs 12.6 [7.1]; P < .001).

Hemoglobin A1c (HbA1c) testing in the previous year was greater than 95% in both groups (P = .77). Complex patients had slightly higher mean HbA1c levels (mean [SD] HbA1c levels, 7.3% [1.5%] vs 6.9% [1.0%]; P = .003) than noncomplex patients, and a larger proportion had at least 1 HbA1c level greater than 9.0% (18.2% vs 8.1%; P = .03) or 10.0% (9.8% vs 1.0%; P = .003) in the previous year. While the proportion of patients who were receiving oral hypoglycemic medications was similar, complex patients were more often prescribed insulin (35.9% vs 19.2%; P = .003). The prevalence of hypertension diagnosis (88.8% complex vs 80.8% noncomplex; P = .06) and treatment (90.6% complex vs 84.6% noncomplex; P = .13) were similar in both groups. Similarly, the rates of lipid-lowering therapy prescription (76.7% vs 70.2%; P = .22) were similar, and good low-density lipoprotein cholesterol control was achieved in both groups (mean [SD] low-density lipoprotein level in previous year, 86.0 [28.5] mg/dL vs 89.4 [26.7] mg/dL [to convert to millimoles per liter, multiply by 0.0259]; P = .34).

Complex patients with diabetes had a significantly higher prevalence of comorbid conditions than noncomplex patients with diabetes. Also, complex patients with diabetes were more likely to be prescribed narcotic analgesics (42.2% vs 25.0%; P = .003), selective serotonin reuptake inhibitors (33.6% vs 14.4%; P < .001), and benzodiazepines (30.0% vs 13.5%; P = .001) and to require physical therapy (36.3% vs 18.3%; P < .001) than noncomplex patients with diabetes. In separate models, after age, sex, and HbA1c levels were controlled for, the likelihood of being designated as complex increased by 2- to 5-fold in the presence of atrial fibrillation (adjusted odds ratio [aOR], 5.0; 95% CI, 2.1-11.7), depression (aOR, 4.6; 95% CI, 1.7-12.2), heart failure (aOR, 3.3; 95% CI, 1.4-8.0), anxiety (aOR, 3.2; 95% CI, 1.3-4.4) osteoarthritis (aOR, 2.4; 95% CI, 1.3-4.4), or chronic obstructive pulmonary disease (aOR, 2.1; 95% CI, 1.1-4.0) (Figure).

Comment

The PCPs who reviewed a randomly selected list of their own patients designated just over two-thirds of their patients with type 2 diabetes mellitus as complex. Prevalence and management of the “core triad” of glycemia, blood pressure, and dyslipidemia did not appear to account for this complexity (with the exception of insulin use), which suggests that dyslipidemia and hypertension do not add substantially to the complexity of patients with type 2 diabetes. These findings indicate that—with the exception of facilitating insulin prescription—we may be reaching a “ceiling” on how much more benefit can be obtained in diabetes-specific quality improvement interventions.

In contrast, significant differences were seen in the prevalence of conditions such as atrial fibrillation, osteoarthritis, depression, and anxiety, conditions that may interfere with either the therapeutic relationship or the self-management activities that are necessary to achieve good diabetes control,4 suggesting that substantial advances in the quality of diabetes primary care may require focusing resources on these other comorbid conditions that complicate diabetes care.

The PCP-identified complex patients had significantly more outpatient clinical care contact during the previous year. Therefore, the patients in our study had ample opportunity to engage with the primary care system. Further improvements in diabetes management may thus require a greater emphasis on comprehensive, patient-centered rather than diabetes-specific interventions.

Back to top
Article Information

Correspondence: Dr Grant, Division of Research, Kaiser Permanente Northern California, 2101 Webster St, Oakland, CA 94612 (Richard.W.Grant@kp.org).

Author Contributions: Dr Grant had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Grant and Atlas. Acquisition of data: Ashburner. Analysis and interpretation of data: Grant, Wexler, Ashburner, and Hong. Drafting of the manuscript: Grant and Ashburner. Critical revision of the manuscript for important intellectual content: Wexler, Ashburner, Hong, and Atlas. Statistical analysis: Wexler and Ashburner. Obtained funding: Grant. Administrative, technical, and material support: Ashburner, Hong, and Atlas. Study supervision: Grant and Atlas.

Financial Disclosure: None reported.

Funding/Support: This study was supported in part by the System Improvement Grant Program, Partners Community HealthCare, Inc.

Additional Contributions: The primary care providers of the Primary Care Operations Improvement Advisory Board, Massachusetts General Hospital, provided clinical insights and valuable feedback during the project.

References
1.
Jackson GL, Edelman D, Weinberger M. Simultaneous control of intermediate diabetes outcomes among Veterans Affairs primary care patients.  J Gen Intern Med. 2006;21(10):1050-1056PubMedArticle
2.
Safford MM, Allison JJ, Kiefe CI. Patient complexity: more than comorbidity: the vector model of complexity.  J Gen Intern Med. 2007;22:(suppl 3)  382-390PubMedArticle
3.
Grant RW, Ashburner JM, Hong CC, Chang Y, Barry MJ, Atlas SJ. Defining patient complexity from the primary care physician's perspective: a cohort study.  Ann Intern Med. 2011;155(12):797-804PubMed
4.
Piette JD, Kerr EA. The impact of comorbid chronic conditions on diabetes care.  Diabetes Care. 2006;29(3):725-731PubMedArticle
×