Piette JD, Mah CA, McPhee SJ; Association for Health Services Research. Meeting.
Abstr Book Assoc Health Serv Res Meet. 1998; 15: 46.
Center for Health Care Evaluation/VA Palo Alto Health Care System and Stanford University, Menlo Park, CA 94025, USA.
RESEARCH OBJECTIVE: Automated telephone disease management (ATDM) may be an effective strategy for monitoring large numbers of diabetic patients. However, providers need evidence that patients' ATDM reports are reliable and clinically meaningful. We examined the reliability and criterion validity of diabetic patients' ATDM-reported self-monitored blood glucose levels (SMBGs), symptoms of hypglycemia and hyperglycemia, foot problems, chest pain, breathing problems, and self-assessments of health and diabetes control. STUDY DESIGN: A total of 235 diabetic patients from both a county medical center and a Veteran's Affairs medical center participated. Patients contributed data from 2,323 successful ATDM calls. A diabetes nurse educator recalled a subset of 78 patients after each of 1,147 ATDM calls, at which time she verified 1,775 problems that patients reported. Nurse callbacks occurred soon afte each atdm call was completed (median of four days). We examined the consistency of reporting within calls (call-level analyses) and the relationship between patients' first ATDM reports and information from their medical records, laboratory tests, and telephone interviews (patient-level analyses). We then used the nurse follow-up data to examine the prevalence and variation in reporting errors across clinical areas (problem-level analyses). PRINCIPLE FINDINGS: Patients had a mean age of 56.2 years, body mass index (BMI) of 33.0 kg/m2, and HgA of 10.1%; 14% of patients spoke Spanish primarily and received all ATDM calls in their native language. Within calls, patients' ATDM reports were consistent. Mean ATDM-reported SMBGs increased from 135 mg/dl for patients reporting "excellent diabetes control" to 215 mg/dl for patients reporting "poor control" (p<.001). Reports of hyperglycemic symptoms increased from 4% when the reported SMBG was <120 mg/dl to 50% when the SMBG was >240 mg/dl (p<.0001). Patients' first ATDM-reported SMBGs were correlated with baseline HgA (r=.5; p=.0001). ATDM-reported fat intake scores correlated with total serum cholesterol (r=.6; p=.003) more than with BMI (r=.2; p=.08). ATDM-reported hypoglycemic symptoms, hyperglycemic symptoms, and SMBGs correlated strongly with interview reports. During nurse follow-up calls, 88% of all ATDM-reported problems (1,562 total problems) were verified by the patient. Patients were more likely to report accurately on general health problems (98% accuracy), glucose control problems (97%), abnormal SMBG levels (93%), and hyperglycemic symptoms (92%) than on hypoglycemic symptoms (72%) and foot problems (gg%; p<.001). Problems reported by younger patients (<65 years) were more likely to be accurate in the areas of hypoglycemic symptoms (76% versus 53%), foot problems (68% versus 46%), and SMBG levels (94% versus 82%; all p<.05). There were no notable differences in the reliability of validity of ATDM reports across groups of patients defined by primary language, education, or severity of illness. CONCLUSIONS: Diabetic patients, including Spanish-speakers and elderly patients, are able to report reliably clinic information using an ATDM system. IMPLICATIONS FOR POLICY, DELIVERY, AND PRACTICE: ATDM could be a useful tool for improving clinical decision-making for diabetes management. ATDM may be especially useful in populations with limited access to health care.
Publication Types:
Keywords:
- Aged
- Data Collection
- Diabetes Mellitus
- Diabetes Mellitus, Type 2
- Disease Management
- Hispanic Americans
- Humans
- Hypoglycemic Agents
- Nurses
- Questionnaires
- Telephone
- therapy
- hsrmtgs
Other ID:
UI: 102234086
From Meeting Abstracts