Agreement between different recall periods in drug utilization studies

Concordância entre diferentes períodos recordatórios em estudos de utilização de medicamentos

Cassia Garcia Moraes Sotero Serrate Mengue Tatiane da Silva Dal Pizzol About the authors

ABSTRACT:

Objective:

To assess the agreement between three recall periods for self-reported drug use using a 24-hour recall period as reference.

Methods:

Participants were allocated into three groups with different recall periods of 7, 14 and 30 days and were interviewed at two different times. A 24-hour recall questionnaire was answered during the first interview, and a questionnaire on drug use over the different recall periods tested was answered during the second interview. The agreement between the questionnaires was evaluated using percent agreement and kappa.

Results:

For continuous drugs, percent agreement varied between 92 and 99% and kappa varied between 0.71 and 0.97 for three periods tested. For drugs of occasional use, percent agreement varied between 63 and 81% and kappa varied between 0.27 and 0.52. The prevalence of drugs, particularly those of occasional use, increases with time.

Conclusions:

The high level of agreement between the three recall periods suggests that all of them are valid for the investigation of drugs of continuous use.

Keywords:
Drug utilization; Questionnaires; Self report; Mental recall; Validation studies; Pharmacoepidemiology

RESUMO:

Objetivo:

Avaliar a concordância entre três períodos recordatórios para o auto-relato do uso de medicamentos, utilizando o período recordatório de 24 horas como referência.

Métodos:

Os participantes foram alocados em três grupos de acordo com os períodos recordatórios de 7, 14 e 30 dias e entrevistados em dois momentos. Um recordatório de 24 horas foi respondido na primeira entrevista e, um questionário sobre o uso de medicamentos com os diferentes períodos recordatórios avaliados foi respondido na segunda entrevista. A concordância entre os questionários foi avaliada pelo percentual de concordância e kappa.

Resultados:

O percentual de concordância variou entre 92 a 99% e o kappa de 0,71 a 0,97 para os medicamentos de uso contínuo nos três períodos recordatórios. Para os medicamentos de uso eventual, o percentual de concordância variou entre 63 a 81% e o kappa oscilou entre 0,27 a 0,52. A prevalência do uso de medicamentos, particularmente para os de uso eventual, aumentou com o tempo.

Conclusão:

A análise de concordância sugere que os três períodos recordatórios avaliados são válidos para a investigação da utilização de medicamentos de uso contínuo.

Palavras-chave:
Uso de medicamentos; Questionários; Autorrelato; Rememoração mental; Estudos de validação; Farmacoepidemiologia

INTRODUCTION

Data were collected exclusively through self-reporting questionnaires, that are either self-administered or administered during face-to-face interviews, used extensively in a variety of health studies, including pharmacoepidemiology studies11. Gama H, Correia S, Lunet N. Effect of questionnaire structure on recall of drug utilization in a population of university students. BMC Med Res Methodol 2009; 9: 45.,22. Pit SW, Byles JE, Cockburn J. Accuracy of Telephone Self-Report of Drug Use in Older People and Agreement with Pharmaceutical Claims Data. Drugs Aging 2008; 25(1). [Internet]. Available from: http://adisonline.com/aging/Fulltext/2008/25010/Accuracy_of_Telephone_Self_Report_of_Drug_Use_in.8.aspx (Access on: December 2nd, 2012).
http://adisonline.com/aging/Fulltext/200...
.

In the majority of cases, the questionnaires are based on recall questions regarding drug use, requiring short, medium or even long-term memory recall from the respondent11. Gama H, Correia S, Lunet N. Effect of questionnaire structure on recall of drug utilization in a population of university students. BMC Med Res Methodol 2009; 9: 45.,33. Klungel OH, de Boer A, Paes AH, Herings RM, Seidell JC, Bakker A. Agreement between self-reported antihypertensive drug use and pharmacy records in a population-based study in The Netherlands. Pharm World Sci PWS 1999; 21(5): 217-20.,44. Haukka J, Suvisaari J, Tuulio-Henriksson A, Lönnqvist J. High concordance between self-reported medication and official prescription database information. Eur J Clin Pharmacol 2007; 63(11): 1069-74.,55. Lunet N, Bastos J, Cumaio F, Silva P, Dias E, Barros H. Recall of drug utilization depends on subtle structural questionnaire characteristics. Pharm World Sci 2008; 30(2): 175-81.. Under-reporting of drug use is considered to result from recall error, even if the error magnitude is not known66. Oliveira EA de, Bertoldi AD, Domingues MR, Santos IS, Barros AJD. Medicine use from birth to age two years: the 2004 Pelotas (Brazil) Birth Cohort study. Rev Saúde Pública 2010; 44(4): 591-600.,77. Santos DB, Barreto ML, Coelho HLL. Drug use and associated factors in children living in poor areas. Rev Saúde Pública 2009; 43(5): 768-78.,88. Arrais PSD, Brito LL, Barreto ML, Coelho HLL. Prevalence and determinants of medicines consumption in Fortaleza, Ceará, Brazil. Cad Saúde Pública 2005; 21(6): 1737-46.,99. Reis A, Tudela M, Jorge P, Novais C, Fonseca C. Prevalence of analgesic consumption in the northern of Portugal. Acta Reumatol Port 2010; 35(5): 434-40.,1010. Ribeiro AQ, Rozenfeld S, Klein CH, César CC, Acurcio FA. Survey on medicine use by elderly retirees in Belo Horizonte, Southeastern Brazil. Rev Saúde Pública 2008; 42(4): 724-32.,1111. Moraes CG, Mengue SS, Tavares NUL, Pizzol D, Silva T. Drug use among children between zero and six years old: a population baseline study in the south of Brazil. Ciênc Amp Saúde Coletiva 2013; 18(12): 3585-93..

The recall period used in studies with self-reported data collection has not yet been standardized1212. Bertoldi AD, Barros AJD, Wagner A, Ross-Degnan D, Hallal PC. A descriptive review of the methodologies used in household surveys on medicine utilization. BMC Health Serv Res 2008; 8: 222.,1313. Costa KS, Francisco PMSB, Barros MBA. Utilização e fontes de obtenção de medicamentos: um estudo de base populacional no Município de Campinas, São Paulo, Brasil. Cad Saúde Pública 2016; 32(1): e00067814.. Each author adopts a preferred period based on convenience or on a trend observed in other studies1313. Costa KS, Francisco PMSB, Barros MBA. Utilização e fontes de obtenção de medicamentos: um estudo de base populacional no Município de Campinas, São Paulo, Brasil. Cad Saúde Pública 2016; 32(1): e00067814.. This practice may hinder the comparison of estimates of drug use prevalence obtained in different studies55. Lunet N, Bastos J, Cumaio F, Silva P, Dias E, Barros H. Recall of drug utilization depends on subtle structural questionnaire characteristics. Pharm World Sci 2008; 30(2): 175-81.,1212. Bertoldi AD, Barros AJD, Wagner A, Ross-Degnan D, Hallal PC. A descriptive review of the methodologies used in household surveys on medicine utilization. BMC Health Serv Res 2008; 8: 222.,1414. Furu K, Thelle DS. Validity of questions in the use of specific drug-groups in health surveys. Pharm World Sci 2001; 23(2): 50-4.,1515. Gama H, Correia S, Lunet N. Questionnaire design and the recall of pharmacological treatments: a systematic review. Pharmacoepidemiol Drug Saf 2009; 18(3): 175-87.. Studies evaluating the recall period and validity of the collected data are scarce. Studies of self-reporting accuracy have used pharmacy and medical prescription records as the gold standard to test the validity of the data1616. Lewis JD, Strom BL, Kimmel SE, Farrar J, Metz DC, Brensinger C, et al. Predictors of recall of over-the-counter and prescription non-steroidal anti-inflammatory drug exposure. Pharmacoepidemiol Drug Saf 2006; 15(1): 39-45.,1717. Drieling RL, LaCroix AZ, Beresford SAA, Boudreau DM, Kooperberg C, Heckbert SR. Validity of Self-Reported Medication Use Compared With Pharmacy Records in a Cohort of Older Women: Findings From the Women's Health Initiative. Am J Epidemiol 2016; 233-8.,1818. Noize P, Bazin F, Pariente A, Dufouil C, Ancelin M-L, Helmer C, et al. Validity of chronic drug exposure presumed from repeated patient interviews varied according to drug class. J Clin Epidemiol 2012; 65(10): 1061-8.. However, such records do not exist in all countries and therefore cannot always be used1212. Bertoldi AD, Barros AJD, Wagner A, Ross-Degnan D, Hallal PC. A descriptive review of the methodologies used in household surveys on medicine utilization. BMC Health Serv Res 2008; 8: 222..

Lewis et al.1616. Lewis JD, Strom BL, Kimmel SE, Farrar J, Metz DC, Brensinger C, et al. Predictors of recall of over-the-counter and prescription non-steroidal anti-inflammatory drug exposure. Pharmacoepidemiol Drug Saf 2006; 15(1): 39-45. studied the influence of the recall period on data accuracy by analyzing the variation in reports of use of prescription and non-prescription drugs over eight weeks, observing a significant decrease in the reported use of the drugs studied, especially drugs of infrequent use and non-prescription drugs.

The objective of this study was to evaluate the level of agreement between 7-, 14- and 30-day recall periods for the self-assessment of drug use using a 24-hour recall period as reference.

METHODS

A cross-sectional study was performed and the sample consisted of professors, staff and students over 18 years of age and capable of communicating from the Universidade Federal do Rio Grande do Sul (UFRGS), located in the south of Brazil. A convenience sampling was performed. Interviewers were instructed to recruit participants by approaching people of different ages, of different types of occupation at the university and of both genders in the hallways, offices and rooms of the University buildings to obtain a heterogeneous sample of the university population. The participants were allocated into the three groups by generating a list of sequential numbers using the software SPSS, version 18.0 (SPSS Inc. Released 2009. PASW Statistic for Windows. Chicago: SPSS Inc.). The three groups consisted of 120 participants each, with three different recall periods frequently used in drug use studies (7, 14 and 30 days) assigned to test the influence of longer and shorter recall periods on the agreement with the 24-hour recall period, adopted as a reference in this study.

During the first interview, the interviewers explained the research procedures and scheduled the date for the second interview according to the allocation group. During the first interview, all participants answered the 24-hour recall period questionnaire. During the second interview, which took place 7, 14 or 30 days following the first meeting, the participants answered a questionnaire on drug use over the past seven, 14 or 30 days, depending on the allocation group. Drug use over the past 7, 14 or 30 days included the 24-hour period that the first interview took place.

Data collection was performed between March and June 2012 using questionnaires with open-ended and closed-ended questions, developed by the authors and previously tested in a pilot study. The outcomes variables were the prevalence of drug use during the last 24 hours, 7, 14 or 30 days, obtained through the question: “Did you use any drugs in the last [24 hours] [7 days] [14 days] [30 days], such as analgesics, continuous medication, vitamins, antihistamines, contraceptive (if a woman), ointments or any other type of medication?”. In case of a positive response, the medication names were inquired. Sociodemographic and medical variables investigated included gender, age, self-reported race/color, family income, occupation at the university (professor, student or staff), levels of education, health insurance plan, number of drugs in the 24 hours and number of drugs in the 7, 14 or 30 days.

For data analysis, drugs were classified into two main groups according to the type of use (continuous or occasional) reported by the participant. Drugs of continuous use were defined as drugs used for the treatment of chronic and/or degenerative diseases, used continuously. Drugs of occasional use were defined as drugs used for the treatment of acute diseases and/or non-prescription drugs. Within these two groups, drugs were distributed among five classes of drugs according to the classes of higher drug use frequency reported in this sample: cardiovascular system and diabetes drugs, female sex hormones, other drugs of continuous use, analgesics and non-steroidal anti-inflammatory drugs (NSAIDs) drugs and other drugs of occasional use. This categorization was adopted so that the prevalence of use in each therapeutic class was pronounced, and therefore would not interfere in the analysis of kappa, which is influenced by low prevalence of the outcome variable. Cardiovascular system and diabetes drugs included anti-hypertensive, lipid-lowering and hypoglycemia drugs. The group “other drugs of continuous use” included anti-asthmatic drugs, anti-thyroid drugs and drugs acting on the central nervous system. Female sex hormones included contraceptive and hormone replacement drugs. The group “other drugs of occasional use” included anti-histamines, antibiotics, expectorants, gastrointestinal and dermatological disorder drugs and supplements, among other categories.

The agreement between the 24-hour recall period and the questionnaires with greater recall periods were evaluated using the kappa coefficient and percent agreement. The level of agreement was based on drugs reported at both the first and second questionnaires. More specifically, for the comparison between the 24-hour recall period and the 7, 14 or 30-day recall periods, the agreement was calculated based on data provided by the 107, 97 or 95 participants who answered the second questionnaire 7, 14 or 30 days after the first questionnaire, respectively. The agreement was evaluated in accordance with the name of the drug. If, for example, in the second questionnaire a participant reported use of “captopril”, there was an agreement if he/she had reported “captopril” in the first questionnaire (24 h). It was included positive agreement (drug reported in the first questionnaire but not in the second) or negative agreement (drug reported in the second questionnaire but not in the first one).

A value of kappa between 0.8 and 1.0 indicated very good agreement, between 0.6 and 0.8 good agreement, between 0.4 and 0.6 moderate agreement, between 0.2 and 0.4 acceptable agreement and lower than 0.2 poor agreement1919. Altman DG. Practical Statistics For Medical Research. Britain: Chapman & Hall; 1991. 628p..

The sociodemographic characteristics of the participants were compared using the Pearson χ2 test for proportions to verify the similarity between groups of the 7, 14 and 30 days recall periods. All analyses were performed using a 5% significance level.

The questionnaires were digitalized using the system Teleform Workgroup V10 and stored and analyzed using the software SPSS, version 18.0.

This study was approved by the Research Ethics Committee of the Universidade Federal do Rio Grande do Sul (UFRGS), under record No. 21,801. Data were collected only after the Informed Consent Form was signed.

RESULTS

Of the 360 participants who answered the 24-hour recall period questionnaire during the first interview, 299 (83%) of them completed the study by answering the second questionnaire. Eleven percent of the participants from the 7-day recall period group, 19% from the 14-day recall period group and 20% from the 30-day recall period group did not answer the second questionnaire and were counted as losses. A significant difference was observed between the participants who completed the study and the ones who did not complete it only for the variable occupation at the university (p < 0.05), with the largest number of losses being observed for category “professors” and for the 30 days recall period group. Not arriving on the scheduled day and place for the second interview and difficulties in contacting the participants were some of the reasons for non-response to the second questionnaire.

The sociodemographic and medical characteristics of the participants are presented in Table 1. The majority of the participants were white women under 50 years old. The average age was 41 years old (SD ± 13.3), and age varied between 18 and 94 years old. The majority of the participants had 12 years or more of education and belonged to the university staff. There were significant differences in family income and occupation (p < 0.05) among the allocation groups. For family income, 7 days recall period group had a lower percentage for income above R$5,000.00 and the group of 14 days, the highest percentage in this category. For occupation, there were fewer professors and more students for 7 days recall period group in relation to the others groups.

Table 1:
Sociodemographic and medical data of study participants in the different allocation groups (n = 299).

The prevalence of drug use for the different allocation groups is presented in Table 2. The prevalence of drugs, particularly those of occasional use, increases with time.

Table 2:
Prevalence of drugs use for the 24-hour and for the 7-, 14- or 30-day recall periods*.

The kappa coefficients and agreement percentages between the 24-hour recall period questionnaire and the questionnaires for the 7-, 14- and 30-day recall periods are presented on Table 3.

Table 3:
Agreement between drug use questionnaires between the 24-hour and the 7-, 14- or 30-day recall periods*.

Agreement between the 24-hour recall period and the longer recall periods tested for drugs of continuous use was high, with values between 92.6% and 99.0%. Kappa was slightly higher for the 7-day recall period group than for the 14- and 30-day groups. The agreement percentages for drugs of occasional use were lower than for drugs of continuous use, between 63.1 and 71.1% for analgesic and NSAIDs and between 72,2 and 81,3% for other drugs of occasional use. Kappa was lower for drugs of occasional use than for drugs of continuous use, between 0.27 and 0.52 for all groups studied.

DISCUSSION

This study investigated the level of agreement of three recall periods of 7, 14 and 30 days through the agreement analysis of questionnaires compared to the 24-hour recall period, in a population sample from a university in the south of Brazil. For drugs of continuous use, the three recall periods tested - 7, 14 and 30 days - presented very good agreement with the reference recall period of 24 hours. For drugs of occasional use, however, the agreement with the reference period (24 hour recall period) was only acceptable.

A small variation in agreement percentage was observed among the 7-, 14- and 30 day recall periods for drugs of continuous use. Kappa was lower for the 14- and 30-day periods than for the 7-day period for all three classes of drugs analyzed. This difference was, however, not very pronounced, indicating that the influence of the error of self-reporting of drug use depending on the recall period adopted is small or null.

The agreement between the questionnaires for the 7-, 14- and 30-day periods and the 24-hour recall period was considered very good, according to Altman1919. Altman DG. Practical Statistics For Medical Research. Britain: Chapman & Hall; 1991. 628p., for cardiovascular system and diabetes acting drugs and for female sex hormones. Agreement was also classified as very good for other drugs of continuous use for the 7- and 14-day periods but not for the 30-day recall period, which was classified as good. These findings indicate a small loss of information among the three recall periods tested and thus a small magnitude of recall error in the reporting of drugs of continuous use.

For analgesic and NSAIDs, Kappa indicated acceptable agreement1919. Altman DG. Practical Statistics For Medical Research. Britain: Chapman & Hall; 1991. 628p. between the three tested recall periods and the 24 hour reference period. An increase in the prevalence of the use of analgesic and NSAIDs was observed for the three recall periods tested when compared with the 24-hour recall period, which was mainly related to the occasional use of these drugs during the different recall periods, which made the analysis of the magnitude of the recall error for these drugs difficult. The different prevalence observed for the different recall periods indicates that the same recall period should be used to allow comparison among studies of the use of analgesic and NSAIDs1616. Lewis JD, Strom BL, Kimmel SE, Farrar J, Metz DC, Brensinger C, et al. Predictors of recall of over-the-counter and prescription non-steroidal anti-inflammatory drug exposure. Pharmacoepidemiol Drug Saf 2006; 15(1): 39-45..

Studies comparing self-reported drug use with pharmacy and medical prescription records reported higher agreement between the pharmacy records and participant reports for drugs of continuous use, such as anti-hypertensive, hormonal therapy, cholesterol lowering, anti-depressant and anti-diabetic drugs than for drugs of occasional use, such as anti-inflammatory drugs. This pattern suggests that the agreement varies with the therapeutic class of the drug1818. Noize P, Bazin F, Pariente A, Dufouil C, Ancelin M-L, Helmer C, et al. Validity of chronic drug exposure presumed from repeated patient interviews varied according to drug class. J Clin Epidemiol 2012; 65(10): 1061-8.,2020. West SL, Savitz DA, Koch G, Strom BL, Guess HA, Hartzema A. Recall accuracy for prescription medications: self-report compared with database information. Am J Epidemiol 1995; 142(10): 1103-12.,2121. Sjahid SI, van der Linden PD, Stricker BH. Agreement between the pharmacy medication history and patient interview for cardiovascular drugs: the Rotterdam elderly study. Br J Clin Pharmacol 1998; 45(6): 591-5.,2222. Klungel OH, de Boer A, Paes AH, Herings RM, Seidell JC, Bakker A. Influence of question structure on the recall of self-reported drug use. J Clin Epidemiol 2000; 53(3): 273-7.,2323. Boudreau DM, Daling JR, Malone KE, Gardner JS, Blough DK, Heckbert SR. A validation study of patient interview data and pharmacy records for antihypertensive, statin, and antidepressant medication use among older women. Am J Epidemiol 200; 159(3): 308-17.,2424. Glintborg B, Hillestrøm PR, Olsen LH, Dalhoff KP, Poulsen HE. Are patients reliable when self-reporting medication use? Validation of structured drug interviews and home visits by drug analysis and prescription data in acutely hospitalized patients. J Clin Pharmacol 2007; 47(11): 1440-9.,2525. Caskie GIL, Willis SL, Warner Schaie K, Zanjani FAK. Congruence of medication information from a brown bag data collection and pharmacy records: findings from the Seattle longitudinal study. Exp Aging Res 2006; 32(1): 79-103.,2626. Lacasse A, Ware MA, Bourgault P, Lanctôt H, Dorais M, Boulanger A, et al. Accuracy of Self-reported Prescribed Analgesic Medication Use: Linkage Between the Quebec Pain Registry and the Quebec Administrative Prescription Claims Databases. Clin J Pain 2016; 32(2): 95-102. and that the high agreement of self-reported drug use found for continuous drugs cannot be generalized to other types of drugs2424. Glintborg B, Hillestrøm PR, Olsen LH, Dalhoff KP, Poulsen HE. Are patients reliable when self-reporting medication use? Validation of structured drug interviews and home visits by drug analysis and prescription data in acutely hospitalized patients. J Clin Pharmacol 2007; 47(11): 1440-9..

Drugs of continuous use to treat chronic diseases, such as cardiovascular system drugs, have been repeatedly observed to be more likely to be accurately reported by individuals than drugs of occasional use, as occasional use drugs are used for self-limiting diseases and for short periods or sporadically2727. Grimaldi-Bensouda L, Rossignol M, Aubrun E, Benichou J, Abenhaim L. Agreement between patients' self-report and physicians' prescriptions on nonsteroidal anti-inflammatory drugs and other drugs used in musculoskeletal disorders: the international Pharmacoepidemiologic General Research eXtension database. Pharmacoepidemiol Drug Saf 2012; 21(7): 753-9.,2828. Caskie GIL, Willis SL. Congruence of self-reported medications with pharmacy prescription records in low-income older adults. Gerontologist 2004; 44(2): 176-85.. The greater ease of study participants in recalling continuous drug use may be related to the daily routine of the use of these drugs2727. Grimaldi-Bensouda L, Rossignol M, Aubrun E, Benichou J, Abenhaim L. Agreement between patients' self-report and physicians' prescriptions on nonsteroidal anti-inflammatory drugs and other drugs used in musculoskeletal disorders: the international Pharmacoepidemiologic General Research eXtension database. Pharmacoepidemiol Drug Saf 2012; 21(7): 753-9.. It should also be considered that the use of drugs of continuous use may be present in the respondent’s memory because daily use is always at a maximal time distance of 24 hours, with the occasional forgetting of one or more drug doses during the reference periods as the only potential recall problem.

Data on drug use obtained directly from the drug user through population surveys is the method most often used in drug use studies in developing countries, including Brazil, where there are no efficient computerized systems that integrate information on drugs and drug users. Such systems are generally considered the gold standard in surveying population drug exposure and are used as references in validity studies22. Pit SW, Byles JE, Cockburn J. Accuracy of Telephone Self-Report of Drug Use in Older People and Agreement with Pharmaceutical Claims Data. Drugs Aging 2008; 25(1). [Internet]. Available from: http://adisonline.com/aging/Fulltext/2008/25010/Accuracy_of_Telephone_Self_Report_of_Drug_Use_in.8.aspx (Access on: December 2nd, 2012).
http://adisonline.com/aging/Fulltext/200...
,1717. Drieling RL, LaCroix AZ, Beresford SAA, Boudreau DM, Kooperberg C, Heckbert SR. Validity of Self-Reported Medication Use Compared With Pharmacy Records in a Cohort of Older Women: Findings From the Women's Health Initiative. Am J Epidemiol 2016; 233-8.,1818. Noize P, Bazin F, Pariente A, Dufouil C, Ancelin M-L, Helmer C, et al. Validity of chronic drug exposure presumed from repeated patient interviews varied according to drug class. J Clin Epidemiol 2012; 65(10): 1061-8.,2323. Boudreau DM, Daling JR, Malone KE, Gardner JS, Blough DK, Heckbert SR. A validation study of patient interview data and pharmacy records for antihypertensive, statin, and antidepressant medication use among older women. Am J Epidemiol 200; 159(3): 308-17.. In the absence of a gold-standard that can be used to test the validity of data on drug use over an optimal recall period, the 24-hour recall period used in this study seems to be a good alternative as the most accurate method for comparison with longer recall periods.

It is necessary to consider that, in addition to factors related to the accurate recall of medication use, there are other factors that may have interfered in the non-agreement between the responses in the different recall periods. For example, changes in responses may represent changes in prescription or pattern of drug use between the periods assessed, or even, by adherence problems in a period of adaptation to medical prescription.

The relatively small sample size, hindering the analysis of therapeutic drug classes with lower use prevalence, and the restriction of the sample to the younger university population, making it relatively socioeconomically homogeneous, are limitations of this study. The choice of the university population was motivated by the easiness location of the individuals, considering the need for a second contact in a determined period of time. For this reason, generalizations of the results of this study for populations much older than the one used in this study, with lower levels of education and family income should be made with caution. Another study limitation was the number of losses, higher for the 14 and 30 day recall periods, where teachers were the majority of those who did not respond to the second questionnaire, mainly in the 30-day recall.

CONCLUSION

The results of this study allow the conclusion that the three recall periods tested (7, 14 and 30 days) seem to have little influence on the estimation of the kappa coefficient for drugs of continuous use, suggesting that the recall error had little effect on the underestimation of the prevalence of the use of drugs of continuous use. As 14- or 15-day recall periods have been used most often in previous studies1313. Costa KS, Francisco PMSB, Barros MBA. Utilização e fontes de obtenção de medicamentos: um estudo de base populacional no Município de Campinas, São Paulo, Brasil. Cad Saúde Pública 2016; 32(1): e00067814.,2929. World Health Organization. EUROHIS: developing common instruments for health surveys. 2003. [Internet]. Available from: http://www.euro.who.int/en/what-we-publish/abstracts/­eurohis-developing-common-instruments-for-health-surveys (Access on: November 13, 2012).
http://www.euro.who.int/en/what-we-publi...
,3030. Cruz MJB, Dourado LFN, Bodevan EC, Andrade RA, Santos DF. Medication use among children 0-14 years old: population baseline study. J Pediatr 2014; 90(6): 608-15.,3131. Tavares NUL, Costa KS, Mengue SS, Vieira MLFP, Malta DC, Júnior S, et al. Uso de medicamentos para tratamento de doenças crônicas não transmissíveis no Brasil: resultados da Pesquisa Nacional de Saúde, 2013. Epidemiol Serviços Saúde 2015; 24(2): 315-23., the adoption of the 14-day recall period is suggested to make comparisons among future studies easier, especially for the investigation of drugs of continuous use. For studies evaluating drugs of occasional use, such as analgesic and non-steroidal anti-inflammatory drugs, we suggest the use of shorter recall periods, in accordance with Lewis et al.1616. Lewis JD, Strom BL, Kimmel SE, Farrar J, Metz DC, Brensinger C, et al. Predictors of recall of over-the-counter and prescription non-steroidal anti-inflammatory drug exposure. Pharmacoepidemiol Drug Saf 2006; 15(1): 39-45..

REFERENCES

  • 1
    Gama H, Correia S, Lunet N. Effect of questionnaire structure on recall of drug utilization in a population of university students. BMC Med Res Methodol 2009; 9: 45.
  • 2
    Pit SW, Byles JE, Cockburn J. Accuracy of Telephone Self-Report of Drug Use in Older People and Agreement with Pharmaceutical Claims Data. Drugs Aging 2008; 25(1). [Internet]. Available from: http://adisonline.com/aging/Fulltext/2008/25010/Accuracy_of_Telephone_Self_Report_of_Drug_Use_in.8.aspx (Access on: December 2nd, 2012).
    » http://adisonline.com/aging/Fulltext/2008/25010/Accuracy_of_Telephone_Self_Report_of_Drug_Use_in.8.aspx
  • 3
    Klungel OH, de Boer A, Paes AH, Herings RM, Seidell JC, Bakker A. Agreement between self-reported antihypertensive drug use and pharmacy records in a population-based study in The Netherlands. Pharm World Sci PWS 1999; 21(5): 217-20.
  • 4
    Haukka J, Suvisaari J, Tuulio-Henriksson A, Lönnqvist J. High concordance between self-reported medication and official prescription database information. Eur J Clin Pharmacol 2007; 63(11): 1069-74.
  • 5
    Lunet N, Bastos J, Cumaio F, Silva P, Dias E, Barros H. Recall of drug utilization depends on subtle structural questionnaire characteristics. Pharm World Sci 2008; 30(2): 175-81.
  • 6
    Oliveira EA de, Bertoldi AD, Domingues MR, Santos IS, Barros AJD. Medicine use from birth to age two years: the 2004 Pelotas (Brazil) Birth Cohort study. Rev Saúde Pública 2010; 44(4): 591-600.
  • 7
    Santos DB, Barreto ML, Coelho HLL. Drug use and associated factors in children living in poor areas. Rev Saúde Pública 2009; 43(5): 768-78.
  • 8
    Arrais PSD, Brito LL, Barreto ML, Coelho HLL. Prevalence and determinants of medicines consumption in Fortaleza, Ceará, Brazil. Cad Saúde Pública 2005; 21(6): 1737-46.
  • 9
    Reis A, Tudela M, Jorge P, Novais C, Fonseca C. Prevalence of analgesic consumption in the northern of Portugal. Acta Reumatol Port 2010; 35(5): 434-40.
  • 10
    Ribeiro AQ, Rozenfeld S, Klein CH, César CC, Acurcio FA. Survey on medicine use by elderly retirees in Belo Horizonte, Southeastern Brazil. Rev Saúde Pública 2008; 42(4): 724-32.
  • 11
    Moraes CG, Mengue SS, Tavares NUL, Pizzol D, Silva T. Drug use among children between zero and six years old: a population baseline study in the south of Brazil. Ciênc Amp Saúde Coletiva 2013; 18(12): 3585-93.
  • 12
    Bertoldi AD, Barros AJD, Wagner A, Ross-Degnan D, Hallal PC. A descriptive review of the methodologies used in household surveys on medicine utilization. BMC Health Serv Res 2008; 8: 222.
  • 13
    Costa KS, Francisco PMSB, Barros MBA. Utilização e fontes de obtenção de medicamentos: um estudo de base populacional no Município de Campinas, São Paulo, Brasil. Cad Saúde Pública 2016; 32(1): e00067814.
  • 14
    Furu K, Thelle DS. Validity of questions in the use of specific drug-groups in health surveys. Pharm World Sci 2001; 23(2): 50-4.
  • 15
    Gama H, Correia S, Lunet N. Questionnaire design and the recall of pharmacological treatments: a systematic review. Pharmacoepidemiol Drug Saf 2009; 18(3): 175-87.
  • 16
    Lewis JD, Strom BL, Kimmel SE, Farrar J, Metz DC, Brensinger C, et al. Predictors of recall of over-the-counter and prescription non-steroidal anti-inflammatory drug exposure. Pharmacoepidemiol Drug Saf 2006; 15(1): 39-45.
  • 17
    Drieling RL, LaCroix AZ, Beresford SAA, Boudreau DM, Kooperberg C, Heckbert SR. Validity of Self-Reported Medication Use Compared With Pharmacy Records in a Cohort of Older Women: Findings From the Women's Health Initiative. Am J Epidemiol 2016; 233-8.
  • 18
    Noize P, Bazin F, Pariente A, Dufouil C, Ancelin M-L, Helmer C, et al. Validity of chronic drug exposure presumed from repeated patient interviews varied according to drug class. J Clin Epidemiol 2012; 65(10): 1061-8.
  • 19
    Altman DG. Practical Statistics For Medical Research. Britain: Chapman & Hall; 1991. 628p.
  • 20
    West SL, Savitz DA, Koch G, Strom BL, Guess HA, Hartzema A. Recall accuracy for prescription medications: self-report compared with database information. Am J Epidemiol 1995; 142(10): 1103-12.
  • 21
    Sjahid SI, van der Linden PD, Stricker BH. Agreement between the pharmacy medication history and patient interview for cardiovascular drugs: the Rotterdam elderly study. Br J Clin Pharmacol 1998; 45(6): 591-5.
  • 22
    Klungel OH, de Boer A, Paes AH, Herings RM, Seidell JC, Bakker A. Influence of question structure on the recall of self-reported drug use. J Clin Epidemiol 2000; 53(3): 273-7.
  • 23
    Boudreau DM, Daling JR, Malone KE, Gardner JS, Blough DK, Heckbert SR. A validation study of patient interview data and pharmacy records for antihypertensive, statin, and antidepressant medication use among older women. Am J Epidemiol 200; 159(3): 308-17.
  • 24
    Glintborg B, Hillestrøm PR, Olsen LH, Dalhoff KP, Poulsen HE. Are patients reliable when self-reporting medication use? Validation of structured drug interviews and home visits by drug analysis and prescription data in acutely hospitalized patients. J Clin Pharmacol 2007; 47(11): 1440-9.
  • 25
    Caskie GIL, Willis SL, Warner Schaie K, Zanjani FAK. Congruence of medication information from a brown bag data collection and pharmacy records: findings from the Seattle longitudinal study. Exp Aging Res 2006; 32(1): 79-103.
  • 26
    Lacasse A, Ware MA, Bourgault P, Lanctôt H, Dorais M, Boulanger A, et al. Accuracy of Self-reported Prescribed Analgesic Medication Use: Linkage Between the Quebec Pain Registry and the Quebec Administrative Prescription Claims Databases. Clin J Pain 2016; 32(2): 95-102.
  • 27
    Grimaldi-Bensouda L, Rossignol M, Aubrun E, Benichou J, Abenhaim L. Agreement between patients' self-report and physicians' prescriptions on nonsteroidal anti-inflammatory drugs and other drugs used in musculoskeletal disorders: the international Pharmacoepidemiologic General Research eXtension database. Pharmacoepidemiol Drug Saf 2012; 21(7): 753-9.
  • 28
    Caskie GIL, Willis SL. Congruence of self-reported medications with pharmacy prescription records in low-income older adults. Gerontologist 2004; 44(2): 176-85.
  • 29
    World Health Organization. EUROHIS: developing common instruments for health surveys. 2003. [Internet]. Available from: http://www.euro.who.int/en/what-we-publish/abstracts/­eurohis-developing-common-instruments-for-health-surveys (Access on: November 13, 2012).
    » http://www.euro.who.int/en/what-we-publish/abstracts/­eurohis-developing-common-instruments-for-health-surveys
  • 30
    Cruz MJB, Dourado LFN, Bodevan EC, Andrade RA, Santos DF. Medication use among children 0-14 years old: population baseline study. J Pediatr 2014; 90(6): 608-15.
  • 31
    Tavares NUL, Costa KS, Mengue SS, Vieira MLFP, Malta DC, Júnior S, et al. Uso de medicamentos para tratamento de doenças crônicas não transmissíveis no Brasil: resultados da Pesquisa Nacional de Saúde, 2013. Epidemiol Serviços Saúde 2015; 24(2): 315-23.

  • Financial support: none

Publication Dates

  • Publication in this collection
    Apr-Jun 2017

History

  • Received
    13 Apr 2016
  • Accepted
    23 Feb 2017
Associação Brasileira de Pós -Graduação em Saúde Coletiva São Paulo - SP - Brazil
E-mail: revbrepi@usp.br