Kathryn R. Koller

Utilizing harmonization and common surveillance methods to consolidate 4 cohorts: the Western Alaska Tribal Collaborative for Health (WATCH) study

Background According to health status reports, chronic disease prevalence appears to be rising in western Alaska Native (AN) people, and accurate population-based data are needed. Four cohort studies of western AN people were conducted in the Norton Sound and Yukon-Kuskokwim regions, but none have been large enough to allow reliable estimates of rates of chronic diseases and evaluate their risk factors. Objective In this article, the methods used to combine 4 major cohort studies of rural western AN people are described and the benefits and challenges encountered in combining data and standardizing surveillance methods for these studies are discussed. Design Tribal permission was obtained for each cohort study and the consolidated study. Data from baseline exams were directly combined or harmonized into new variables. Common surveillance methods were developed and implemented to identify incidence and risk factors for cardiovascular disease (CVD) events and type 2 diabetes. Results A cohort of 4,569 western AN participants (2,116 men and 2,453 women), aged 18–95 years, was established to study CVD and diabetes prevalence. Prospective surveillance data over an average 6.7-year follow-up can now be used to study CVD and diabetes incidence and associated risk factors in a subset of 2,754 western AN participants (1,218 men and 1,536 women) who consented to initial surveillance. Conclusions The combined cohort provides statistical power to examine incidence rates and risk factors for CVD and diabetes and allows for analyses by geographic region. The data can be used to develop intervention programmes in these populations and others.

A comparison of in-person examination and video otoscope imaging for tympanostomy tube follow-up.

The objective of this study was to determine if video otoscope still images (640 x 480 pixel resolution) of the tympanic membrane following surgical placement of tympanostomy tubes are comparable to an in-person microscopic examination. Forty patients having undergone tympanostomy tube placement in both ears were independently examined in-person by two otolaryngologists and imaged using a video otoscope and telemedicine software package. The two physicians later reviewed images at 6 and 12 weeks. Physical examination findings and diagnosis were documented and compared for their concordance using kappa statistics. For both physicians, the intraprovider concordance between the in-person examination and the corresponding image review was high for each of the physical examination findings: Tube In 93-94% (K 0.85-0.87), Tube Patent 86-93% (K 0.74-0.85), Drainage 94-98% (K 0.42-0.66), Perforation 85-98% (K 0.40-0.84), Granulation 95-99% (K -0.01 to 0.00), Middle Ear Fluid 89-91% (K -0.03 to 0.50), and Retracted 89-94% (K 0.13-0.43). These agreement rates are similar to the normal interprovider concordance observed when two physicians independently examined the same patient in-person for physical exam findings: Tube In 96% (K 0.93), Tube Patent 94% (K 0.88), Drainage 96% (K 0.56), Perforation 90% (K 0.60), Granulation 96% (K 0.39), Middle Ear Fluid 88% (K 0.14), and Retracted 91% (K 0.43). For both physicians, the intraprovider diagnostic concordance between the in-person examination and the corresponding image review was high 79-85% (K 0.67-0.76). The interprovider diagnostic concordance for the in-person exam was 88% (K 0.81). The interprovider diagnostic concordance when two physicians independently reviewed all images was 84% (K 0.74), and 89% (K 0.80) when poor images were excluded. This study demonstrates that physician review of video otoscope images is comparable to an in-person microscopic examination. Store-and-forward video otoscopy may be an acceptable method of following patients post-tympanostomy tube placement.

Digital images for postsurgical follow-up of tympanostomy tubes in remote Alaska

Objective To determine if video otoscope still images of the tympanic membrane taken in remote clinics are comparable to an in-person microscopic examination for follow-up care. Design Comparative concordance, diagnostic reliability. Methods Community health aide/practitioners in remote Alaska imaged 70 ears following tympanostomy tube placement. The patients were then examined in person by two otolaryngologists. Images were later reviewed at 8 and 14 weeks. Results Intraprovider concordance for physical examination findings was: “Tube in,” 94 percent-97 percent (κ = 0.89-0.94); “Tube patent,” 94 percent −97 percent (κ = 0.89-0.94); “Drainage,” 90 percent-96 percent (κ = –0.04-0.38); “Perforation,” 90 percent −96 percent (κ = 0.61-0.82); “Granulation,” 97 percent −100 percent (κ = 0.49-1.0); “Middle ear fluid,” 88 percent −96 percent (κ = 0.28-0.71); “Retracted,” 83 percent-91 percent (κ = 0.26-0.58). These agreement rates are similar to interprovider concordance when two otolaryngologists examine the same patient in person. Intraprovider concordance for diagnoses was 76 percent −80 percent (κ = 0.64-0.71) and 77 percent −88 percent (κ = 0.66-0.81) when poor images were excluded. Interprovider diagnostic concordance for the in-person exam was 89 percent (κ = 0.83). Conclusion Video-otoscopy images of the tympanic membrane are comparable to an in-person examination for assessment and treatment of patients following tympanostomy tubes. Store-and-forward telemedicine is an acceptable method of following patients post tympanostomy tube placement.

Breast cancer among Alaska Native women potentially exposed to environmental organochlorine chemicals

Objectives. To determine if an increased rate of breast cancer in Alaska Native women is related to their consumption of a subsistence diet that may contain p, p’-dichlorodiphenylethylene (DDE) and polychlorinated biphenyls (PCBs). Study Design. A retrospective case control design. Methods. We analyzed banked serum collected between 1981 and 1987 from 126 Alaska Native women, including 63 case women who subsequently developed breast cancer and 63 age-matched control women who remained cancer-free. Serum was analyzed for DDT, DDE, 13 other chlorinated pesticides, and 28 PCB congeners. Results. The geometric mean for p, p’-DDE levels among case women was 8.67 ppb (95% Confidence Interval 7.48, 10.04); among control women, the geometric mean was 7.36 ppb (6.53, 8.30). The geometric mean for total PCB levels among case women was 4.55 ppb (3.61, 5.74) and for control women, the geometric mean was 6.10 ppb (4.73, 7.86). Cancer status and total PCB levels varied across ethnicity (i.e., Eskimo, Aleut, and Indian) but DDE levels were uniform among these ethnic groups. Using conditional logistic regression analysis to adjust for potential confounders (e.g., ethnicity, family history of breast cancer, parity), we found an odds ratio of 1.43 (0.46, 4.47) for the highest tertile of DDE exposure and 0.42 (0.07, 2.38) for the highest tertile of total PCB exposure. Conclusions. Although the results are limited by small sample size and restricted risk factor information, our findings of higher DDE levels, but lower PCB levels among women with breast cancer are consistent with previous research. Our results confirm exposure to organochlorines among Alaska Native women but do not identify these exposures as a significant risk factor for breast cancer.

Case–control study of breast cancer and exposure to synthetic environmental chemicals among Alaska Native women

Background Exposure to environmental chemicals may impair endocrine system function. Alaska Native (AN) women may be at higher risk of exposure to these endocrine disrupting chemicals, which may contribute to breast cancer in this population. Objective To measure the association between exposure to select environmental chemicals and breast cancer among AN women. Design A case-control study of 170 women (75 cases, 95 controls) recruited from the AN Medical Center from 1999 to 2002. Participants provided urine and serum samples. Serum was analyzed for 9 persistent pesticides, 34 polychlorinated biphenyl (PCB) congeners, and 8 polybrominated diethyl ether (PBDE) congeners. Urine was analyzed for 10 phthalate metabolites. We calculated geometric means (GM) and compared cases and controls using logistic regression. Results Serum concentrations of most pesticides and 3 indicator PCB congeners (PCB-138/158; PCB-153, PCB-180) were lower in case women than controls. BDE-47 was significantly higher in case women (GM=38.8 ng/g lipid) than controls (GM=25.1 ng/g lipid) (p=0.04). Persistent pesticides, PCBs, and most phthalate metabolites were not associated with case status in univariate logistic regression. The odds of being a case were higher for those with urinary mono-(2-ethylhexyl) phthalate (MEHP) concentrations that were above the median; this relationship was seen in both univariate (OR 2.16, 95% CI 1.16-4.05, p=0.02) and multivariable (OR 2.43, 95% CI 1.13-5.25, p=0.02) logistic regression. Women with oestrogen receptor (ER)-/progesterone receptor (PR)-tumour types tended to have higher concentrations of persistent pesticides than did ER+/PR+ women, although these differences were not statistically significant. Conclusions Exposure to the parent compound of the phthalate metabolite MEHP may be associated with breast cancer. However, our study is limited by small sample size and an inability to control for the confounding effects of body mass index. The association between BDE-47 and breast cancer warrants further investigation.Background Exposure to environmental chemicals may impair endocrine system function. Alaska Native (AN) women may be at higher risk of exposure to these endocrine disrupting chemicals, which may contribute to breast cancer in this population. Objective To measure the association between exposure to select environmental chemicals and breast cancer among AN women. Design A case–control study of 170 women (75 cases, 95 controls) recruited from the AN Medical Center from 1999 to 2002. Participants provided urine and serum samples. Serum was analyzed for 9 persistent pesticides, 34 polychlorinated biphenyl (PCB) congeners, and 8 polybrominated diethyl ether (PBDE) congeners. Urine was analyzed for 10 phthalate metabolites. We calculated geometric means (GM) and compared cases and controls using logistic regression. Results Serum concentrations of most pesticides and 3 indicator PCB congeners (PCB-138/158; PCB-153, PCB-180) were lower in case women than controls. BDE-47 was significantly higher in case women (GM=38.8 ng/g lipid) than controls (GM=25.1 ng/g lipid) (p=0.04). Persistent pesticides, PCBs, and most phthalate metabolites were not associated with case status in univariate logistic regression. The odds of being a case were higher for those with urinary mono-(2-ethylhexyl) phthalate (MEHP) concentrations that were above the median; this relationship was seen in both univariate (OR 2.16, 95% CI 1.16–4.05, p=0.02) and multivariable (OR 2.43, 95% CI 1.13–5.25, p=0.02) logistic regression. Women with oestrogen receptor (ER)–/progesterone receptor (PR)-tumour types tended to have higher concentrations of persistent pesticides than did ER+/PR+ women, although these differences were not statistically significant. Conclusions Exposure to the parent compound of the phthalate metabolite MEHP may be associated with breast cancer. However, our study is limited by small sample size and an inability to control for the confounding effects of body mass index. The association between BDE-47 and breast cancer warrants further investigation.

Diabetes prevalence, incidence, complications and mortality among Alaska Native people 1985–2006

Objectives. To examine trends in diabetes prevalence, incidence, complications and mortality between 1985 and 2006 among Alaska Native people. Study design. We used data from the population-based Alaska Native Diabetes Registry, which includes all people who receive care in the Alaska Tribal Health System. Methods. We compared the periods of 1986–1990 and 2002–2006 for diabetes-related amputations, renal replacement and mortality using Poisson regression. Complications and mortality data were examined for trends using Poisson regression. Survival analyses for those diagnosed since 31 December 1985 were performed using the Cox proportional hazard model. Results. Age-adjusted diabetes prevalence increased from 17.3 in 1985 to 47.6/1,000 in 2006. The number of Alaska Native people living in Alaska with diabetes increased from 610 in 1985 to 3,386 in 2006. Diabetes incidence rates have also increased. Comparing age-adjusted rates for the 5-year periods 1986–1990 and 2002–2006, amputations decreased from 5.3 to 2.6/1,000, renal replacement decreased from 3.3 to 1.2/1,000 and mortality decreased from 41.7 to 33.2/1,000. Yearly analyses showed a downward trend for amputations, renal replacement and mortality rates. Survival analyses showed a significantly higher hazard ratio for any amputations, major amputations and renal replacement for the earlier time period compared to the most recent time period. Conclusions. An increase in risk factors, awareness, funding and case-finding may be contributing to the increase in prevalence and incidence of diagnosed diabetes. While diabetes prevalence and incidence are increasing among Alaska Native people, our results suggest that even in remote, rural areas, complications and mortality can be reduced.

Effect of special diabetes program for Indians funding on system changes in diabetes care and outcomes among American Indian/Alaska Native people 1994-2004.

OBJECTIVES: The Alaska Native Medical Center diabetes program analysed Diabetes Care and Outcomes Audit data from 1994–2004 to evaluate the impact of the Special Diabetes Program for Indians (SDPI) funding on process and intermediate outcomes. STUDY DESIGN: We conducted a retrospective analysis of data from standardized medical records reviews conducted between 1994 and 2004 from regional sites in Alaska. METHODS: We analysed 7,735 randomly selected records for trends over three time periods (pre-SDPI, transition and SDPI). RESULTS: Hemoglobin A1c, total and LDL cholesterol, triglycerides and blood pressure significantly improved from the pre-SDPI to the SDPI period. However, as the number of people with diabetes increased, the percentage of patients receiving foot, eye and dental exams decreased, as did the percentage receiving nutrition, exercise and diabetes education. CONCLUSIONS: SDPI funding provided resources for interventions necessary to improve the effectiveness of diabetes care. This was associated with improved intermediate outcomes in American Indian/Alaska Native patients with diabetes. Further observations are needed to evaluate whether or not intermediate outcomes result in decreased cardiovascular disease, amputations, dialysis and retinopathy.

Performance on rheumatoid arthritis quality indicators in an Alaska Native healthcare system

Objective Rheumatoid arthritis (RA) is highly prevalent in some Alaska Native and American Indian populations. Quality indicators for RA have been proposed, but these have not been widely implemented or used to assess RA care in Alaska Native or American Indian populations. Methods Medical records were included if they met the following criteria: RA diagnosed before October 2000 fulfilling American College of Rheumatology classification criteria; all care for RA at the Alaska Native Medical Center. Records were reviewed for a 5-year period to determine compliance with eight quality indicators defined by the Arthritis Foundation Quality Indicator Program. Multivariate logistic regression was performed to analyse associations with quality of care. Results There were 106 individuals included in the study. The highest-scoring measures were folic acid prescription if on methotrexate (93.3%) and disease-modifying antirheumatic drug prescription (90.6%). The lowest scoring measure was radiographs of both hands and feet (16.0%). In multivariate analysis, the factor most strongly associated with disease-modifying antirheumatic drug prescriptions, annual exam for RA and hand radiographs was at least one visit with a rheumatologist. Conclusions Quality of care for RA varies by measure and is better for patients who see a rheumatologist. These data provide an initial evaluation of RA quality of care in a unique minority population with an integrated healthcare system.

Frequency of defective DNA mismatch repair in colorectal cancer among the Alaska Native people

The frequency of colorectal cancer (CRC) among the Alaska Native people is the highest of any ethnic group in the United States. In this study, CRC from 329 Alaska Native people (165 Eskimo, 111 Indians, and 53 Aleut) were evaluated for evidence of defective DNA mismatch repair (MMR) by testing tumors for altered protein expression (hMLH1, hMSH2, and hMSH6) and for the presence of microsatellite instability. Of the 329 samples tested, 46 (14%) showed both microsatellite instability and altered protein expression; 42 (91%) with a loss of hMLH1, 3 (7%) hMSH2, and 1 (2%) hMLH1/hMSH6. Tumors with loss of hMLH1 were further evaluated for hMLH1 promoter hypermethylation and for the presence of the BRAF-V600E mutation. Among cases tested, all 19 (100%) tumors among the Eskimo patients showed hMLH1 promoter hypermethylation, whereas this was the case for only 3 of the 7 (42%) tumors among the Aleut (P = 0.002) and 5 of the 10 (50%) tumors from the Indians (P = 0.002). The majority of hypermethylated cases (23 of 27) tested positive for the V600E alteration. Of the nine tumors from the Aleut and Indian patients that did not exhibit hMLH1 hypermethylation, eight tested negative for V600E. Overall, the frequency of defective MMR among the three groups was not statistically different (P = 0.75). However, the data suggest that the pathogenesis of CRC may differ between the three groups. The CRC with defective MMR among the Eskimo sample fit the typical profile for hMLH1-related cancer associated with sporadic CRC, whereas the pattern in the Aleut and Indian suggests the possibility that germ line hMLH1 mutations may be present. (Cancer Epidemiol Biomarkers Prev 2007;16(11):2344–50)

Cardiometabolic correlates of low type 2 diabetes incidence in western Alaska Native people - The WATCH study.

AIMS Previously rare among Alaska Native (AN) people, type 2 diabetes (DM2) prevalence as indicated by registry data has increased by as much as 300% in some western Alaska regions. We sought to determine prevalence and incidence of DM2 and analyze associated cardiometabolic risk factors in western AN people. METHODS DM2 and prediabetes prevalence and incidence were determined by the Western Alaska Tribal Collaborative for Health using consolidated data from cohort studies conducted during 2000-2010. Crude and age-adjusted incidence for DM2 and prediabetes were calculated using 2010 American Diabetes Association criteria. Effects of covariates on DM2 and prediabetes were determined using univariate and multivariate Cox proportional hazards analyses, adjusted for age and sex. RESULTS Excluding baseline diabetes (n=124, 4.5%), 53 cases of new DM2 were identified among 2630 participants. Age- and sex-adjusted DM2 incidence was 4.3/1000 (95% CI 2.9, 5.0) person-years over an average 5.9-year follow up. After excluding baseline prediabetes, 387 new cases of prediabetes were identified among 1841 participants; adjusted prediabetes incidence was 44.5/1000 (95% CI 39.5, 49.5) person years. Independent predictors for DM2 included age, impaired fasting glucose, and metabolic syndrome; family history of diabetes and obesity were additional independent predictors for prediabetes. CONCLUSIONS DM2 incidence in western AN people is substantially lower than that for U.S. whites; however, incidence of prediabetes is more than 10-fold higher than western AN DM2 incidence and more closely aligned with U.S. rates. Interventions aimed at achieving healthy lifestyles are needed to minimize risk factors and maximize protective factors for DM2 in this population.