|
|
|
|
Practical Information
|
Instrument Name:
|
General Adherence Scale (GAS)
|
Instrument Description:
|
This 5-item measure assesses patients’ general tendencies and their likelihood that they will remain on prescribed medical treatment.
|
Price:
|
Free: items are available in the literature. (Ref: 1-3)
|
Administration Time:
|
Less than 5 minutes
|
Publication Year:
|
1986 (First used in the Medical Outcomes Study)
|
Item Readability:
|
Flesch-Kincaid Grade Level of 6.5 for the items. Most items included less than 15 words and were written in simple language.
|
Scale Format:
|
5-item, 6-point Likert scale
|
Administration Technique:
|
No information found.
|
Scoring and Interpretation:
|
Responses are averaged after reversing the scoring of items 1 and 3, and transformed to a linear 0 – 100 distribution. (Ref: 1) A higher score means a higher degree of adherence to medical recommendations.
|
Forms:
|
15-item Specific Adherence Scale (Medication, Exercise, and Diet) was reported. (Ref: 1-3)
|
|
|
Research Contacts
|
Instrument Developers:
|
M. Robin DiMatteo, PhD, Ron D. Hays, PhD and Cathy D. Sherbourne, PhD and others in the Medical Outcomes Study conducted by RAND.
|
Instrument Development Location:
|
M. Robin DiMatteo
Department of Psychology
University of California
Riverside, California 92521-0426
Ron D. Hays
RAND Health Communications
1700 Main Street P.O. Box 2138
Santa Monica, CA 90407-2138
|
Instrument Developer Email:
|
robin.dimatteo@ucr.edu
|
Instrument Developer Website:
|
www.rand.org/health/surveys/mos.patientadh.html
|
|
|
Annotated Bibliography
|
1. DiMatteo MR, Hays RD, Sherbourne CD. Adherence to cancer regimens: implications for treating the older patient. Oncology (Huntingt). 1992 Feb;6(2 Suppl):50-7. [PMID: 1532737]
 |
Purpose: To describe six factors that affect adherence to cancer regimens in older population, and to examine the difficulties that medical professionals face when enhancing the adherence of their patients. |
|
Sample: The data from the Medical Outcome Study were used here. Analyses were based on 2181 patients, and four age groups were recognized. Age group 18-44 consisted of 779 people, 45-64 consisted of 843 people, 65-74 consisted of 656 people, and there were 266 people in the 75+ age group. |
|
Methods: In the Medical Outcomes Study, a general and a specific adherence scale were developed and tested to patients with heart disease, diabetes, and hypertension twice over a 2-year time interval. |
|
Implications: While there had been no consensus relationship between a patient’s age and his/her degree of adherence, the authors suggested that an older patient had more difficulties to adhere to medical treatment than a younger patient. |
2. DiMatteo MR, Sherbourne CD, Hays RD, Ordway L, Kravitz RL, McGlynn EA, Kaplan S, Rogers WH. Physicians' characteristics influence patients' adherence to medical treatment: results from the Medical Outcomes Study. Health Psychol. 1993 Mar;12(2):93-102. [PMID: 8500445]
|
Purpose: To assess the degree of the influence of physicians’ own personal characteristics and their ways of practices on patient adherence. |
|
Sample: The data from Medical Outcomes Study (MOS) were used in this study. Non-psychiatric physicians: N = 186. The characteristics of the physicians are described as follows: Mean age = 40.1 years (SD = 6.7), White = 82%, Practice Type (Solo = 54%, Large multispeciality group = 17%, HMO = 29%); Patients: N = 2546 |
|
Methods: The data were collected at the six points in the MOS study. 1) Patients Screening Self-Report, 2) Tracer Condition Phone Interview, 3) Patients Assessment Questionnaire – Baseline, 4) Medical History Questionnaire – Baseline, 5) Patient Assessment Questionnaire – Year Two, and 6) Physician Report Forms. |
|
Implications: Physicians’ personal characteristics were not found to affect on patient adherence, but physicians’ global job satisfaction affected on patients’ general adherence. |
3. Hays RD. The medical outcomes study (MOS) measures of patient adherence. [serial online]. Available from: http://www.rand.org/health/surveys_tools/mos/mos_adherence.html. Accessed 2002 Aug 21.
|
Purpose: To describe two adherence scales, the general and the specific adherence scales, used in the Medical Outcomes Study. |
|
|
top
|
Factors and Norms
|
Factor Analysis Work:
|
Yes, but no detailed information was found. (Ref: 2)
|
Normative Information Availability:
|
No normative data were found, though means were reported for large samples from the MOS study. These mean scores and standard deviations (SD) at baseline for four age groups were obtained from all medical patients in the MOS study who completed the GAS included. (Ref: 1)
| Age Group
|
Sample Size
|
Mean (SD)
|
| 18-44
|
779
|
64.55 (23.02)
|
| 45-64
|
843
|
72.56 (22.01)
|
| 64-74
|
656
|
78.61 (19.63)
|
| 75 and older
|
266
|
81.53 (19.53)
|
|
|
|
Reliability Evidence
|
Test-retest:
|
Test-retest reliability for 2 years was 0.39. (Ref: 2) 0.40 for 2 years was also reported. (Ref: 1, 3)
|
Inter-rater:
|
No information found.
|
Internal Consistency:
|
Internal consistency reliabilities were reported as 0.81 at baseline for two studies (Ref: 1,3) and .78 in another study (Ref: 2). At a 2-year follow up, internal consistency was reported as 0.79. (Ref: 2)
|
Alternate Forms:
|
No information found.
|
|
|
Validity Evidence
|
Construct/ Convergent/ Discriminant:
|
Authors report that factor analysis and multitrait scaling analysis supported the unidimensionality of the five items, however, few details regarding these analyses are presented. (Ref: 2) The correlations between the general adherence scale and the specific adherence scales ranged from –0.12 to 0.29. (Ref: 3) In a study with 2,181 medical patients, correlations between the self-reported GAS and objective indicators of adherence in a cancer trial were -.13 for GAS and blood glucose levels, -0.24 for GAS and body mass index, and -0.11 with a dyspnea scale. All of the correlations were statistically significant with p < .05.
|
Criterion-related/ Concurrent/ Predictive:
|
The baseline GAS significantly predicted the GAS at 2 years when more than 30 patient and physician variables were controlled in a study with patients from 186 physicians. The standardized beta weight was 0.27. (Ref: 2)
|
Content:
|
No information found.
|
Responsiveness Evidence:
|
No information found.
|
Scale Application in VA Populations:
|
No information found.
|
Scale Application in non-VA Populations:
|
Yes. (Ref: 1-2)
|
|
|
Comments
|
The General Adherence Scale is a short measure of patients’ perceptions of their tendency to adhere to physician recommendations. Since the measure is short, it can easily be included in clinical practice or research studies. An advantage of the measure is that it was developed and tested in a large sample of patients. A limitation of the measure is that the development sample was predominantly white. Furthermore, the GAS has not been tested in veteran sample, to our knowledge.
The means and standard deviations for the age groups were not reported by medical condition. Comparison of means of future samples to this sample should take this into consideration. The validity evidence of the GAS was not strong. The correlations between the GAS and self-report measures of specific adherence were quite low. Furthermore, the correlations between the GAS and what were considered objective measures of adherence were also small (less than 0.25). These low correlations might indicate poor support for the validity of the GAS, but could just as easily indicate that the measures of objective adherence were poor measures. Stronger validity evidence would be obtained by comparing the GAS to better measures of adherence, such as pill counts or urine or blood assay of medication in the patient’s body. Since adherence is thought to be a good thing, developers of the GAS suggest that a response bias might influence scores. The correlation between the GAS and a measures of social desirability was 0.15, providing evidence that the bias is not very strong. Further research that gathers stronger validity evidence and examines possible response bias in the GAS would be beneficial.
One additional note: Table 5 in reference 1 presents rates of adherence for specific physician recommendations by patient subgroups. The authors use a cut-off of “All of the time” as signifying adherence versus non-adherence. This decision rule is not justified in the article and appears inconsistent with their other use of both general and specific adherence (i.e., as continuous measures).
|
Updates
No information found.
|
|
|
