Document Type



This report describes a field-based study that was used to implement a community-based health screening program. The program was designed to determine temporal changes in the prevalence of obesity, non-insulin dependent diabetes mellitus (NIDDM), and other health measures on the Navajo Reservation. The Many Farms-Rough Rock community was selected for this screening program because it had previously been involved in a multi-disciplinary community health project conducted by the Cornell Medical School in 1955 through 1961. The project combines field data collection with standard laboratory techniques.The sampling area was approximately 900 miles of rough terrain with few good roads. There were no accurate maps or rosters of community residents. Few households have telephone or mail delivery. The project team relied on direct field recruitment and data collection in order to minimize selection bias and to maximize participation. The services of an interpreter were essential, since many adults speak Navajo exclusively or as their language of choice. A geographically-based cluster of sampling design was selected for the project. This sampling design is frequently used by the World Health Organization in areas with unreliable or non-existent population rosters. The sample covers approximately 30 sites, with examination of an average of 7 individuals from each site. All adults 20 years or older were eligible. Blood samples were necessary to determine random blood glucose (RBG). A non-invasive technique (finger stick capillary sampling) was thought to be more acceptable to subjects as opposed to venipuncture. This method also presented fewer technical difficulties to the survey team. First, a single drop of capillary blood was analyzed with a portable glucometer to provide immediate analysis. A Capillary Blood Collector with serum separator gel was used to collect an additional 8-10 drops of blood for later lab analysis. The collectors were stored in a cooler for 30 minutes and then spun on a compact, lightweight, high-speed centrifuge. The centrifuge was powered by a compact 12v, 110w power-inverter which plugged into the truck\'s cigarette lighter. The blood samples were then stored in the cooler until the end of the day. They were taken to lab personnel at the local Indian Health Service facility (Chinle). The samples were analyzed on a spectrophotometric analyzer using the coupled hexokinase procedure. Individuals with an RBG of greater than or equal to 7.8 mmol/e, were contacted for a follow-up fasting glucose tolerance test at Chinle. Subjects were transported to Chinle by the team\'s interpreter. Of the 29 individuals contacted for a follow-up glucose tolerance test, only one refused to be evaluated. Data was collected from 231 out of 242 individuals contacted.The 95% response rate is considerably higher than that of similar community-based health surveys in which the subjects are evaluated in a clinic rather than in their residential locations. Only eleven of those contacted refused evaluation, two of whom were known diabetics. Accepting the validity of self-reporting of the diabetes status, only 4 (18%) of the 22 known diabetics responded that they did not have NIDDM. No one inaccurately identified themselves as being diabetic. In evaluating the effectiveness of the portable glucometer in the field situation, the results revealed a correspondence with lab results in 44% of the instances charting hyperglycemia values, and correspondence with 99% for instances of normoglycemic values. If the portable glucometer was the only test, then more than half of the ""true positives"" would not have been called back for further testing. The inaccuracies of the portable glucometer were attributed to extremes of heat and reduced oxygen tension.The findings of this study strongly suggest that a combination of field and standard lab techniques are necessary to ensure both high subject participation and reliable test results.

Publication Date



Indian Health Service, Staff Office of Planning, Evaluation, and Research, Rockville, MD 20857 (E-107).