Comparison of Self-Reported and Physician-Reported Antidepressant Medication Use

doi:10.1016/S1047-2797(98)00072-6

Annals of Epidemiology

Volume 9, Issue 5, July 1999, Pages 283-289

https://doi.org/10.1016/S1047-2797(98)00072-6 Get rights and content

Abstract

PURPOSE: Self-reported medication histories obtained in pharmacoepidemiologic case-control studies are subject to non-differential misclassification and to recall bias. The accuracy of self-reported antidepressant medication use has never been evaluated, but it is important in light of the hypothesis that antidepressant medications may be associated with cancer risk.

METHODS: Within a case-control study of several cancer sites, we compared self-reported antidepressant medication use with antidepressant use recorded in physicians’ records. All female cases (n = 147) and controls (n = 119) who reported antidepressant medication use, and a 10% random sample (n = 114) of those who reported no antidepressant use, were asked to provide consent to contact, and the name(s) of their physician(s). These physicians completed a data abstraction form including information on antidepressant prescriptions recorded in patients’ medical records.

RESULTS: Substantial agreement was found between subject- and physician-reported antidepressant medication use (kappa = 0.60 [95% confidence interval (CI), 0.47–0.74]; agreement = 80%), and use of specific antidepressant medications (agreement ranged from 82 to 100%), while moderate agreement was observed for duration of use (weighted kappa = 0.56 (95% CI, 0.32–0.79)), and date of first use [weighted kappa = 0.48 (95% CI, 0.23–0.72)]. The level of agreement did not differ markedly between cases and controls, except for duration of use, where agreement was somewhat greater for cases.

CONCLUSIONS: The similar level of agreement among cases and controls suggests that differential misclassification (e.g., recall bias) is unlikely in the reporting of most aspects of antidepressant medication use by women. Furthermore, the overall accurate self-reporting of antidepressant use suggests that there should be minimal non-differential antidepressant exposure misclassification.

Introduction

Self-reported medication histories obtained in pharmacoepidemiologic case-control studies are subject to non-differential misclassification and to recall bias. While previous studies have evaluated the accuracy of self-reported use of oral contraceptives and estrogen replacement therapy, few have examined non-hormonal medication use. Generally, recall of hormonal drug use was quite accurate, and recall of non-hormonal medication use poor 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13. The accuracy of self-reported antidepressant medication use is not known. However, it is important because many studies use self-reported antidepressant medication history. For example, epidemiological studies have evaluated the hypothesis that antidepressants may be associated with cancer risk 14, 15, 16, and the hypothesis that antidepressant use maybe associated with age at menopause (17). Studies have also been conducted to describe temporal trends in antidepressant medication use 18, 19.

Misclassification of antidepressant exposure may bias relative risk estimates associated with antidepressant use, and may result in inaccurate prevalence estimates.

The present study compared antidepressant medication use reported by female cancer cases and controls participating in the Health Canada Enhanced Cancer Surveillance (ECS) study, with prescription information documented in physicians’ medical records, in an attempt to assess the potential for misclassification of antidepressant exposure and recall bias.

Section snippets

Data Source and Data Collection

Female cases and controls were sampled from the Ontario component of the ECS study, a case-control study designed to evaluate the association between environmental factors and various cancers. The ECS study identified cancer cases, aged 20–74 years and diagnosed during 1995–96, using the Ontario Cancer Registry. Controls were randomly sampled from the population-based assessment rolls of the Ontario Ministry of Finance and were frequency matched, within 5-year age groupings, to the cases. Data

Results

Non-participants did not differ markedly from participating women with respect to most characteristics (Table 1), suggesting that response bias was minimal. Age group, education level, clinical depression, cancer site, and current smoking status were not associated with subject participation. Cases were more likely to participate than controls, as were subjects with higher household income. A slightly lower proportion of participants reported antidepressant use as compared to those not

Discussion

This study found substantial agreement between subject- and physician-reported `ever’ antidepressant medication use, and antidepressant medication names, while moderate agreement was observed for duration of use, dose, and date of first use. The degree of agreement did not differ markedly between cases and controls, except in regard to duration, where agreement was somewhat greater for cases than controls.

Similar studies of other non-hormonal medications reported poorer recall accuracy.

Acknowledgements

This research was performed within the context of the Enhanced Cancer Surveillance project, sponsored by the Laboratory Centre for Disease Control, Health Canada. Michelle Cotterchio is a research student of the National Cancer Institute of Canada supported with funds provided by the Canadian Cancer Society. We thank members of the Ontario Enhanced Cancer Surveillance Steering Committee and the study staff of ECS, particularly Bonnie James, the Senior Project Manager. We also thank Drs. Yang

References (30)

A. Coulter et al.
The ability of women to recall their oral contraceptive histories
Contraception.
(1986)
A. Steingart et al.
Do antidepressants cause, promote or inhibit cancer?
J Clin Epidem.
(1995)
S.L. West et al.
Completeness of prescription recording in outpatient medical records from a health maintenance organization
J Clin Epidemiol.
(1994)
S.L. West et al.
Demographics, health behaviors, and past drug use as predictors of recall accuracy for previous prescription medication use
J Clin Epidemiol.
(1997)
H.C. Kraemer et al.
Kappa coefficients in epidemiologyan appraisal of a reappraisal
J Clin Epidemiol.
(1988)
W.D. Thompson et al.
A reappraisal of the Kappa coefficient
J Clin Epidemiol.
(1988)
A. Paganini-Hill et al.
Reliability of recall of drug usage and other health-related information
Am J Epidemiol.
(1982)
S.A. Adam et al.
A case-control study of the possible association between oral contraceptives and malignant melanoma
Br J Cancer.
(1981)
P.A. Van den Brandt et al.
Comparison of questionnaire information and pharmacy data on drug use
Pharmaceutisch Weekblad Scientific.
(1991)
S. Harlow et al.
Agreement between questionnaire data and medical records
Am J Epidemiol.
(1989)

S.L. West et al.

Recall accuracy for prescription medicationsSelf-report compared with database information

Am J Epidemiol.

(1995)

R. Glass et al.

Accuracy of recall of histories of oral contraceptive use

Br J Prev Soc Med.

(1974)

P.D. Stolley et al.

Agreement rates between oral contraceptive users and prescribers in relation to drug use histories

Am J Epidemiol.

(1978)

M.J. Rosenberg et al.

Agreement between women’s histories of oral contraceptive use and physician records

Int J Epidemiol.

(1983)

P. Nischan et al.

Comparison of recalled and validated oral contraceptive histories

Am J Epidemiol.

(1993)

Cited by (48)

The international prevalence of antidepressant use before, during, and after pregnancy: A systematic review and meta-analysis of timing, type of prescriptions and geographical variability
2020, Journal of Affective Disorders
Citation Excerpt :
When we stratified by definition of SSRI use, we found a higher prevalence estimate for studies using pharmacy records (prescription/dispensing data) compared to studies relying on self-report. There is some evidence that self-reported psychiatric medication use is less accurate (Haapea et al., 2010; Van den Brandt et al., 1991), as a result of social desirability bias or self-stigmatization (Cotterchio et al., 1999; Nielsen et al., 2008; Rauma et al., 2013), but a recent large population-based study showed the opposite: a very good agreement between antidepressant self-report and prescription data (Hafferty et al., 2018). The observed difference might therefore rather reflect a difference in included study population.
Antidepressant use during pregnancy has increased over the last decades, while safety has been under debate. Our aim was to measure the international prevalence of antidepressant use before, during, and after pregnancy and examine timing, type of prescriptions and geographic variability.
We searched Embase, Medline Ovid, Web of Science, Cochrane Central and Google Scholar from their inception until February 19, 2019. We determined pooled prevalence estimates of antidepressants before, during, and after pregnancy, as well as stratified according to substantive variables.
We identified 40 cohorts from 15 countries, together reporting on 14,072,251 pregnancies. Included studies had a low risk of bias, often reporting on large representative cohorts. Selective serotonin reuptake inhibitors (SSRIs) were the most commonly used antidepressants during pregnancy, with an international prevalence estimate of 3.0% (95%CI 2.3;3.7). While Europe and Australasia had pooled prevalence estimates of 1.6% and 1.3% respectively, Northern America had a prevalence estimate of 5.5% (Q-value = 126.19; df = 2; p-value<0.01). Highest SSRI prevalence rates were found for sertraline (1.10%), followed by citalopram and fluoxetine (0.77% and 0.76% respectively) (Q-value = 121.25; df = 5; p-value<0.01). Qualitative analysis indicated an increase in antidepressant use over subsequent calendar years.
Substantial heterogeneity remained unaccounted for throughout the analyses, even after accounting for hypothetical contributors.
This meta-analysis revealed substantial regional differences in antidepressant use around pregnancy, which could be due to variability in prescription behavior, healthcare seeking behavior and organization of healthcare. There is an urgent need for evidence on effectiveness, benefit, and harm of antidepressants during pregnancy to guide clinical practice.
Self-reported medication use validated through record linkage to national prescribing data
2018, Journal of Clinical Epidemiology
Citation Excerpt :
Self-report can be compromised by a number of factors, including not understanding the question, poor recall, and intended nondisclosure [4]. There is no consensus on patient-level factors predisposing to discordance between medication self-report and gold standard measures, but previous reports have implicated advancing age [9,19], being unmarried [19,21], number of medications regularly dispensed [18,22], suffering poor health [19], and lower educational attainment [21]. Within medication classes, there is some evidence that psychiatric medications are less likely to be accurately self-reported [19,22].
Researchers need to be confident about the reliability of epidemiologic studies that quantify medication use through self-report. Some evidence suggests that psychiatric medications are systemically under-reported. Modern record linkage enables validation of self-report with national prescribing data as gold standard. Here, we investigated the validity of medication self-report for multiple medication types.
Participants in the Generation Scotland population-based cohort (N = 10,244) recruited 2009–2011 self-reported regular usage of several commonly prescribed medication classes. This was matched against Scottish NHS prescriptions data using 3- and 6-month fixed time windows. Potential predictors of discordant self-report, including general intelligence and psychological distress, were studied via multivariable logistic regression.
Antidepressants self-report showed very good agreement (κ = 0.85, [95% confidence interval (CI) 0.84–0.87]), comparable to antihypertensives (κ = 0.90 [CI 0.89–0.91]). Self-report of mood stabilizers showed moderate-poor agreement (κ = 0.42 [CI 0.33–0.50]). Relevant past medical history was the strongest predictor of self-report sensitivity, whereas general intelligence was not predictive.
In this large population-based study, we found self-report validity varied among medication classes, with no simple relationship between psychiatric medication and under-reporting. History of indicated illness predicted more accurate self-report, for both psychiatric and nonpsychiatric medications. Although other patient-level factors influenced self-report for some medications, none predicted greater accuracy across all medications studied.
The association between antidepressant use and hemoglobin A1C in older adults
2017, Geriatric Nursing
Citation Excerpt :
This study was also limited by cross-sectional design, length of time between exposure and outcome measurement and use of self-reported medication data. However, previous research has found self-reported antidepressant medication data to agree with medication use data collected from other sources.26,27 One study performed in an older sample found moderate agreement between report of antidepressant medications taken in the past 6 months and pharmacy records with less agreement for medications taken within the last 2 or 8 years.28
Depression is known to increase diabetes risk and worsen glycemic control in older adults, who already experience high rates of diabetes. The independent impact of antidepressants on glucose control is less clear. Data was drawn from the Health and Retirement Study, a large nationally-representative longitudinal study of retired individuals. Crude and adjusted linear models stratified by diabetes status were used to examine the cross-sectional associations between antidepressant use categorized by subclass and continuous hemoglobin A1C. The sample included 1,153 individuals, most over the age of 70. Antidepressant use was not associated with hemoglobin A1C in any model whether stratified or in the total combined sample. Antidepressants as a class were also not associated with hemoglobin A1C. These findings add to the literature suggesting that antidepressants are not associated with diabetes risk or glycemic control. Prospective studies with larger sample sizes are needed to confirm this finding.
Patterns, predictors, and patient-reported reasons for antidepressant discontinuation in the WHO World Mental Health Surveys
2023, Psychological Medicine
Impact of depressive symptoms on medication adherence in older adults with chronic neurological diseases
2024, BMC Psychiatry
Accuracy of a self-report prescription opioid use diary for patients discharge from the emergency department with acute pain: a multicentre prospective cohort study
2022, BMJ Open

View all citing articles on Scopus

View full text

Original reportsComparison of Self-Reported and Physician-Reported Antidepressant Medication Use

Abstract

Introduction

Section snippets

Data Source and Data Collection

Results

Discussion

Acknowledgements

Contraception.

J Clin Epidem.

J Clin Epidemiol.

J Clin Epidemiol.

J Clin Epidemiol.

J Clin Epidemiol.

Reliability of recall of drug usage and other health-related information

Am J Epidemiol.

A case-control study of the possible association between oral contraceptives and malignant melanoma

Br J Cancer.

Comparison of questionnaire information and pharmacy data on drug use

Pharmaceutisch Weekblad Scientific.

Agreement between questionnaire data and medical records

Am J Epidemiol.

Recall accuracy for prescription medicationsSelf-report compared with database information

Am J Epidemiol.

Accuracy of recall of histories of oral contraceptive use

Br J Prev Soc Med.

Agreement rates between oral contraceptive users and prescribers in relation to drug use histories

Am J Epidemiol.

Agreement between women’s histories of oral contraceptive use and physician records

Int J Epidemiol.

Comparison of recalled and validated oral contraceptive histories

Am J Epidemiol.

Original reports
Comparison of Self-Reported and Physician-Reported Antidepressant Medication Use