The study was conducted at a University-affiliated Children's Hospital (the Center for Health and Technology [CHT] at the University of California Davis Medical Center [UCDMC]). California covers approximately 175,000 square miles and is the third largest and the most populous state in the union. Approximately 90% of the California land mass is considered rural.32
According to 2000 Census data, rural California includes 5 million residents or 14.8% of the total population of California, and this population is increasing.33
Since the establishment of the telemedicine program at UCDMC in 1996, the CHT has provided nearly 17,000 live interactive videoconference medical consultations in more than 40 specialties to more than 80 rural and remote sites in California. Eighteen percent of these consultations have been provided to children and adolescents.
Office of Human Research Protection
This research protocol was reviewed and approved by the Office of Human Research Protection at the UCDMC prior to the initiation of this study.
Telemedicine weight management consultations provided to children and adolescents 18 years of age and younger were identified using the UCDMC Telemedicine Clinical Consultation Database. Between January 2000 and September 2004, 139 children and adolescents at 18 rural clinics in California received such consultations from a UCDMC pediatric weight management specialist.
The Telemedicine Clinical Consultation Database was used to determine patient age, gender, International Classification of Diseases (ICD) 9th Edition diagnoses codes, length of time from referral to telemedicine consultation, and number of days in the telemedicine treatment program. Our analyses were restricted to patients referred for diagnoses of obesity, morbid obesity, overweight, or abnormal weight gain (ICD 9th Edition classification codes 278.0, 278.01, 278.02, or 783.1, respectively). Patient medical records were reviewed by either a pediatric weight management specialist not involved with any of the telemedicine consultations, or by a research associate at the UCDMC.
Telemedicine consultations consisted primarily of one-on-one patient evaluations provided by a weight management specialist and/or endocrinologist with a rural healthcare provider present at the remote site. All consultations were initiated by primary care providers at the rural clinics. Consultations were performed in accordance with the UCD Health System telemedicine policies and procedures.34
Video consultations were performed at 384 kilobits per second by either Internet Protocol (IP) or Integrated Services Digital Network (ISDN).35,36
A retrospective review of patient medical records was conducted to determine whether outpatient telemedicine weight management consultations for obese children and adolescents resulted in (1) changes or additions to patient diagnoses, (2) changes or additions to diagnostic evaluation, (3) changes or additions to treatment, and (4) improvement in patient diet (increased consumption of fruits and/or vegetables, reduced consumption of sugar sweetened beverages, reduced fat consumption, or more structured meal patterns), activity level (increased physical activity and/or reduced sedentary activities), and weight status (weight maintenance, weight loss, or slowing of rate of weight gain). Coding guidelines for the first three outcomes were based on a previously published study conducted by researchers at the UCDMC-CHT, which examined the effectiveness of telemedicine consultations in dermatology, endocrinology, and psychiatry.37
Changes or Additions to Patient Diagnoses
Referral forms and patient charts were reviewed to determine whether there were changes made to patient diagnoses, or whether additional diagnoses related to co-morbidities of obesity were made by the consultant (for example, dyslipidemia, acanthosis nigricans, insulin resistance, type 2 diabetes, hypertension, depression, obstructive sleep apnea, and asthma). We coded this information as (1) changes/additions to diagnosis, (2) no changes/additions to diagnosis, or (3) diagnosis pending test results, as best as can be determined from chart. A more specific diagnosis made by the consultant was coded as a “change” (for example, “snoring” versus “obstructive sleep apnea”). If the referring diagnosis from the referring provider did not include the consultant's diagnosis, it was coded it as a “change.” No change in diagnosis was coded if the consultant confirmed the referring provider's diagnosis and did not add any other diagnosis relevant to obesity. If the consultant's note did not include any specific diagnosis due to pending (laboratory, radiological, or other) test results, we coded it as “diagnosis pending test results.” In most cases this category was not used because there was typically at least one diagnosis that was added or changed.
Changes or Additions to Diagnostic Evaluation
Notes from the consulting provider were reviewed to determine changes or additions to the diagnostic evaluation. A change or addition to diagnostic evaluation was coded if the consultant modified or added to laboratory, radiological or other test plans. Referral to other consultants (for example psychiatry, pulmonology, endocrinology or cardiology) was also coded as a change or addition to evaluation, as was screening the patient for psychosocial issues like depression or anxiety, using self-administered questionnaires. No change or addition to diagnostic evaluation was noted when no other tests or subspecialist referrals were recommended by the consultant.
Changes or Additions to Treatment
Similarly, notes from the consultant were reviewed to determine whether any changes were made to treatment. A change in treatment was documented if there were any changes made by the consultant to the patient's medication regimen, including initiation of new medications, discontinuation of current medications, or modifications of doses of current medications. This included changes in over-the-counter as well as prescription medications. We also coded changes or additions to treatment if the consultant documented behavior modification strategies utilized at the visit or made recommendations for referral to subspecialty services such as endocrinology, cardiology, pulmonology, or psychiatry. We coded no changes or additions to treatment if no medications were added, modified, or discontinued.
Improvement in Diet, Activity, and Weight Status
Only patients who had two or more visits documented in their charts were included in analyses for nutrition, activity, and clinical improvement. Improvement was measured by comparing clinical outcomes between the first and last visits for changes in patient dietary habit, activity level, and weight status. Improvement in diet was coded if the consultant documented increased consumption of fruits and/or vegetables, reduced consumption of sugar-sweetened beverages, reduced fat consumption, or more structured meal patterns. Improvement in activity level was coded if the consultant documented increased physical activity and/or reduced sedentary activity. Improvement in weight status was coded if the consultant documented or the reviewer noted weight maintenance, weight loss, or slowing of rate of weight gain between the first and last visits. If there was no note by the consultant of diet or activity level improvement or if the consultant or the reviewer did not document improvement in weight status, it was coded as “no improvement.”
Two reviewers reviewed a randomly selected subset of eligible patient charts to evaluate coding reliability, accuracy, and interrater reliability. The first reviewer was an attending physician at the face-to-face weight management clinic at UCDMC who was not involved in the telemedicine consultations included in this study. The second reviewer was an experienced telemedicine research assistant who had reviewed patient charts in a published study with similar outcomes in other telemedicine specialty clinics. Each reviewer independently abstracted information on all outcome variables. A reabstraction was conducted on 23% of the charts to measure interrater reliability using a Kappa statistic.
Odds ratios (ORs) and 95% confidence intervals (CIs) were computed to determine the association between clinical improvement in patient diet, activity and weight status, and changes or additions to diagnoses, diagnostic evaluation, and treatment. Interrater reliability was calculated using the Kappa statistic. Statistical analyses were performed using SPSS 14.0 for Windows © 1989–2005 (SPSS Inc., Chicago, IL).