Impact of chronic comorbid illnesses on diabetes care among veterans with type 2 diabetes mellitus
Citation & Export
Hide
Simple citation
Pentakota, Sri Ram.
Impact of chronic comorbid illnesses on diabetes care among veterans with type 2 diabetes mellitus. Retrieved from
https://doi.org/doi:10.7282/T3BZ642PExport
Description
TitleImpact of chronic comorbid illnesses on diabetes care among veterans with type 2 diabetes mellitus
Date Created2013
Other Date2013-10 (degree)
Extentix, 136 p. : ill.
DescriptionClinical management of diabetes is complicated by chronic comorbidities. We examined the impact of close to 60 comorbidities on diabetes care, which were categorized into 5 chronic comorbid illness groups (CCIGs) based on degree of overlap/relatedness of their management plans with those for diabetes (Piette and Kerr framework): none (no comorbidity), concordant (very related, e.g. cardiovascular diseases), discordant (unrelated, e.g. mental/musculoskeletal illnesses), both concordant and discordant, and dominant (can overwhelm care priorities, e.g. metastatic cancer). We conducted 5 separate analyses on cohorts of veterans with recent-onset diabetes who sought care at Veterans Health Administration (VHA) facilities nation-wide (95% male; average age 66 years) and hypothesized that having discordant/dominant illnesses would be associated with poor diabetes care, rapid rise in mean HbA1c levels following initial drop after treatment initiation for diabetes, lower adherence/persistence with diabetes medication regimens, and lower treatment intensification following diabetes treatment failure. Concordant illnesses presence was hypothesized to have limited or a positive impact on diabetes care. We first evaluated association between CCIGs and diabetes care in a cohort of 42, 826 veterans with new onset diabetes in fiscal year (FY) 2003. Study outcomes were 5 guideline measures (HbA1c and LDL-C testing, diabetes-related visits, HbA1c < 8%, and LDL-C < 130 mg/dL) assessed in FY2004. Those with concordant illnesses received similar or better diabetes care compared to those with no comorbidities. Those with discordant and dominant illnesses received poorer care (odds lower by 10-21% and 32-54%, respectively). The second analysis followed 79,249 veterans who initiated diabetes oral mono-therapy in FY2000-02 until they were censored at either the end of FY2010, death (28.6%), or lost to follow-up (5.8%). We compared HbA1c trends (992,196 tests) using piecewise linear random effects models. The models compared 3 HbA1c parameters: initial drop for first 6 months, HbA1c at end of 6 months, subsequent rise till end of study. Rate of rise of mean HbA1c was steeper for none (0.071%/year) and discordant CCIGs (0.081%/year). Rise in mean HbA1c for veterans with concordant (0.055%/year) and dominant (0.052%/year) illnesses were more moderate. The third and fourth analyses (n=79,249) compared medication adherence (proportion of days covered (PDC) =>0.8) and non-persistence (treatment gap => 60 days) across CCIGs using logistic and Cox proportional hazards models, respectively. Half the cohort had PDC =>0.8. Discordant and dominant illnesses lowered odds for adherence by 12-32%. Dominant CCIG was more likely to be non-persistent [hazard ratio (95% CI)-1.12 (1.08-1.17)]. The fifth analysis utilized a cohort of 28, 472 veterans who failed initial diabetes treatment (first HbA1c > 8%). Treatment intensification (addition of second oral agent or insulin initiation) rates within 1 year of treatment failure were compared across CCIGs. Concordant and dominant CCIGs were associated with lower treatment intensification odds (by 10%). In a large well-integrated managed care organization like VHA with limited access barriers, we found mixed support for the Piette and Kerr framework. Dominant illnesses lowered HbA1c regardless of inferior care processes, supporting need for patient-specific treatment goals to avoid complications from hypoglycemia. This questions the validity of HbA1c as quality measure for that group. Veterans with no or discordant comorbidities were associated with relatively lower adherence and poor maintenance of glycemic control, representing need for intervention via better care coordination. Strengths include large, population-based study with high prevalence of comorbidity. Key limitation is results not generalizable to the U.S. population.
NotePh.D.
NoteIncludes bibliographical references
Noteby Sri Ram Pentakota
Genretheses, ETD doctoral
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Statistical Profile