Skip to main content
Download PDF

Asking informed consent may lead to significant participation bias and suboptimal cardiovascular risk management in learning healthcare systems

BMC Medical Research Methodology volume 23, Article number: 98 (2023) Cite this article

Abstract

Background

The Utrecht Cardiovascular Cohort – CardioVascular Risk Management (UCC-CVRM) was set up as a learning healthcare system (LHS), aiming at guideline based cardiovascular risk factor measurement in all patients in routine clinical care. However, not all patients provided informed consent, which may lead to participation bias. We aimed to study participation bias in a LHS by assessing differences in and completeness of cardiovascular risk management (CVRM) indicators in electronic health records (EHRs) of consenting, non-consenting, and non-responding patients, using the UCC-CVRM as an example.

Methods

All patients visiting the University Medical Center Utrecht for first time evaluation of a(n) (a)symptomatic vascular disease or condition were invited to participate. Routine care data was collected in the EHR and an informed consent was asked. Differences in patient characteristics were compared between consent groups. We performed multivariable logistic regression to identify determinants of non-consent. We used multinomial regression for an exploratory analysis for the determinants of non-response. Presence of CVRM indicators were compared between consent groups. A waiver (19/641) was obtained from our ethics committee.

Results

Out of 5730 patients invited, 2378 were consenting, 1907 non-consenting, and 1445 non-responding. Non-consent was related to young and old age, lower education level, lower BMI, physical activity and haemoglobin levels, higher heartrate, cardiovascular disease history and absence of proteinuria. Non-response increased with young and old age, higher education level, physical activity, HbA1c and decreased with lower levels of haemoglobin, BMI, and systolic blood pressure. Presence of CVRM indicators was 5–30% lower in non-consenting patients and even lower in non-responding patients, compared to consenting patients. Non-consent and non-response varied across specialisms.

Conclusions

A traditional informed consent procedure in a LHS may lead to participation bias and potentially to suboptimal CVRM, which is detrimental for feedback on findings in a LHS. This underlines the importance of reassessing the informed consent procedure in a LHS.

Peer Review reports

Introduction

As cardiovascular diseases are the leading cause of death globally, international cardiovascular risk management (CVRM) guidelines have been developed to reduce cardiovascular morbidity and mortality [1,2,3,4]. These guidelines are meant to assist health professionals in the decision making process and include cardiovascular risk indicators that should be assessed in all at-risk patients, such as, smoking status, blood pressure, serum lipids, as well as recommendations for treatment of these risk factors. However, compliance to assessment of cardiovascular risk indicators according to these guidelines varies considerably between treating specialists [1, 5]. Therefore, the Center for Circulatory Health of the University Medical Center (UMC) Utrecht initiated the cross-specialty Utrecht Cardiovascular Cohort-CardioVascular Risk Management (UCC-CVRM) project in 2015. UCC-CVRM aims for hospital-wide uniform assessment and registration of guideline based cardiovascular risk indicators in all patients referred for cardiovascular evaluation in a learning healthcare system (LHS) approach [6]. In addition, broad informed consent was asked for scientific purposes: reuse of healthcare data, including linkage, for scientific research and blood sample storage in a biobank [6, 7].

A LHS is a cycle in which “evidence informs practice, and practice informs evidence” [8]. In other words; routine clinical care data collected from the EHR is analysed and the knowledge generated from these analyses is used to change healthcare practices by, for example, closing the feedback loop to healthcare professionals [8]. An important characteristic of a LHS is its inclusivity, all patients visiting a participating department of the LHS participate [9]. However, when a traditional informed consent procedure is in place, which was the case in UCC-CVRM, only data from consenting patients is processed and used to create new evidence and improve the quality of care. We previously showed that not all UCC-CVRM patients provided informed consent, 57%, potentially impairing the possible benefits of the approach in terms of complete registration and assessment of cardiovascular risk status in all patients [9].

Non-consent is an old but persisting phenomenon with causes ranging from forgetting to return the consent form (more often categorised as “non-response”) to actively stating not wanting to participate at all (more often categorised as “non-consent”) [10]. The clinical significance and consequence of non-consent and non-response may be a misrepresentation of the actual population by only those who provided informed consent, called participation bias [11,12,13]. Under Dutch law, non-consenting and non-responding patients belong to the same group ‘patients that were asked, yet did not provide written informed consent’. To date, however, because of missing informed consents, information on the characteristics of non-consenting and non-responding individuals within clinical research is limited and, when available, mostly limited to examination of age and sex distributions [12]. Yet, participation bias may lead to under- or overestimation of associations and hampers generalisations and conclusions [14].

Therefore, the aims of this study were to gain insight into the differences between consenting, non-consenting and non-responding patients and the determinants thereof. Furthermore, we aimed to assess the differences in structured registration of cardiovascular risk indicators between the consent groups.

Methods

We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational (cross-sectional) studies. The STROBE statement can be found in Supplement 1 Table 1 (S1T1).

Study setting

The rationale and design of the ongoing prospective UCC-CVRM cohort study have been described elsewhere [6]. In short, all patients visiting one of the Center for Circulatory Health’s departments at the UMC Utrecht for the first time to evaluate a(n) (a)symptomatic vascular disease or condition were invited to participate. All eligible patients received the invitation via regular mail, including information and the UCC-CVRM questionnaire. Patients filled out a questionnaire on demographics, intoxications, medical history, medication use, physical activity, family cardiovascular disease (CVD) history and pregnancy history before their planned visit, or afterwards, in case of emergency visits. The full questionnaire was added as supplemental material (S2). Additionally, a set of data, based on the recommendations in the Dutch CVRM guidelines, was collected in the EHR as part of routine clinical care. Prior to the appointment with the medical specialist, patients were seen by a trained research nurse to answer any remaining questions regarding the questionnaire. Then, patients were asked if they would consider participation in the UCC-CVRM. Written informed consent was requested for the use of routine clinical care data for scientific research, blood sample storage in the biobank, and linkage with registries for follow-up [6]. All information, including routine care data, questionnaire data and informed consent data were made available through the UMC Utrecht Research Data Platform. All participating departments in the LHS received monthly progress reports about the number of invited patients in the previous month and the percentage of patients that provided written informed consent. In addition, of the patients with a written informed consent, the percentage with a UCC-CVRM questionnaire registered in the EHR and registered laboratory measurements was shown, and the percentage for which we were able to calculate the 10-years risk of CVD morbidity or mortality. The present dataset was based on information collected from May 2015 to December 2019.

Study population

We restricted the analyses to patients aged 18 years or older with planned visits at an outpatient clinic. We excluded specialties with less than 10 patients in either the consent, non-consent, or non-response group. This was the case for the vascular surgery, cardiothoracic surgery, infectious diseases, obstetrics and the multidisciplinary cardiovascular department. Besides, the obstetrics and gynaecology specialisms were excluded because we wanted to study sex-differences and including these departments would lead to an overrepresentation of women. The neurology department did not have an outpatient clinic and was, therefore, excluded. Additionally, only consenting or non-consenting patients up until December 31st 2019 were included in the analyses because of complete change in health care during the COVID-19 pandemic in 2020 [15]. The non-response group consisted of patients that were invited to participate in the UCC-CVRM, but never responded to the invitation. For them the invitation date was used to select the patients up until December 31st, 2019.

Variable source and definition

Sex, blood pressure, BMI, heart rate and laboratory measurements were extracted from structured fields in the EHR. For blood pressure the measurement ± 7 days of the UCC-CVRM inclusion date was used, for all other measurements the value closest to inclusion date ± 21 days. Measurements outside of these cut-offs were not considered clinically relevant for the patient’s visit to the UMC Utrecht. In these cases, a missing value was recorded. Age was calculated at inclusion. Self-reported data from the UCC-CVRM questionnaire was used to obtain information on education level, smoking status, cardiovascular history, and other cardiovascular risk factors. Hyperlipidemia was defined as having an LDL-cholesterol level higher than 3 mmol/L. We defined ‘high education’ as having obtained a degree from a University of Applied Sciences or general University. Furthermore, patients were asked if they were diagnosed with diabetes, excluding gestational diabetes. Total activity (METminutes) per week was calculated using self-reported data from the validated SQUASH questionnaire [16]. Absolute 10-year risk scores for cardiovascular disease (CVD) morbidity and mortality were calculated for each patient using prevailing algorithms at the time of the start of UCC-CVRM, being either the SCORE risk model (SCORE-NL) [17], United Kingdom Prospective Diabetes Study risk score (UKPDS) [18], or the Second Manifestations of Arterial Disease (SMART) score [19], depending on the patient’s medical history.

Data analyses

We presented the patient characteristics in strata of consent, non-consent, and non-response in number and percentages, as means and standard deviations or medians and the first and third quartiles, as appropriate. Additionally, we presented the sex distribution within age categories and treating specialties, stratified by consent status.

Due to the nature and large amount of missing data in the non-response group, we decided to exclude them from the regression analyses exploring determinants of non-consent, and thus proceeded with presentation of two groups only (see S3T1). Potential determinants of non-consent were selected based on literature and findings of descriptive statistics. For the analyses regarding the department, we took cardiology as a reference since most patients were recruited from that department. For age, we took the 70–79 year group as a reference since that was the age group with the largest number of invited subjects. Multiple imputation was used to deal with missing values before fitting the multivariable logistic regression model, using the MICE package in R [20]. To avoid biased imputations, variables were imputed separately by consent status. We used backward selection to fit the logistic regression model and pooled the estimates. Assumptions of logistic regression were tested prior to the analyses, e.g., continuous determinants were categorized in case of a non-linear relationship with the outcome. To explore the non-responding group in more detail, we also imputed that group, and repeated the determinants analyses using multinomial regression analyses [21, 22].

Finally, we assessed the clinical implications of non-consent and non-response, studying CVRM measurement extractability in the EHR stratified by consent status as a proxy for the health care professional’s compliance to the CVRM guidelines. The data was considered extractable if it was registered in the designated field of the EHR.

R version 4.0.5. was used for all analyses [23].

Results

Patient characteristics

Out of the 5730 patients that were invited for participation between May 2015 and December 2019 (S4F1), 41.5% were consenting (N = 2378), 33.3% non-consenting (N = 1907) and 25.2% non-responding (N = 1445) (Table 1). Non-consenting patients were, compared to consenting patients, older, more often women, referred to the geriatric department, and had a higher cardiovascular burden: higher systolic blood pressure, higher 10-year CVD morbidity or mortality risk and more often a cardiovascular disease history. Non-responding patients were more physically active, had lower lipid values, less often a cardiovascular disease history and a lower 10-year CVD risk than both consenting and non-consenting patients. But please note that in the non-responding group, missing information was very considerable and limits the validity of these comparisons using the crude data.

Table 1 General characteristics stratified by consent status, based on the information extractable from the EHR
Full size table

Age and treating specialty stratified by sex

The non-consent group had the highest percentage of patients aged 70 years or older (54.3%) compared to 33.4% of the consent-group and 43.0% of the non-response group. In the non-consent group, most patients in the lowest and the two highest age groups were women (69%, 59% and 78%, respectively, Fig. 1). We performed sensitivity analyses to investigate whether the differences in the age distribution between the consent, non-consent and non-response group could be related to the high number of patients referred to the geriatric department. These analyses illustrated that the geriatrics department had less variability in age than the other departments and that their age was considerable higher (S5F1). Analyses excluding the geriatric department showed that the age differences between the consent and non-consent group declined (S5T1).

Fig. 1
figure 1

Sex distribution within age categories across consent groups

Full size image

The percentage of consenting, non-consenting, and non-responding patients varied across medical specialties (Fig. 2). Geriatrics accounted for almost half of all non-consenting patients (47.7%). Geriatrics and vascular medicine were least equally distributed in terms of sex in the non-consent group (18% more women than men).

Fig. 2
figure 2

Sex distribution within treating specialties across consent groups

Full size image

Determinants of non-consent

Results from the multivariable logistic regression model showed that patients with increased odds of non-consent were being referred to the geriatric department (OR 2.13 (1.79–2.55)), aged between 30–39 or 80–89 years old, more often had a heartrate > 100 bpm (OR 2.06 (1.42–2.98)), HbA1c levels > 48 mmol/mol (OR 1.35 (1.08–1.70)) and a history of cardiovascular diseases (OR 1.43 (1.23–1.66)) (Fig. 3, S6T1). Furthermore, patients with lower odds of non-consent had a high education level (OR 0.76 (0.60–0.97)), proteinuria (OR 0.69 (0.55–0.87)), and higher BMI and physical activity levels. Additionally, being referred to the vascular medicine, nephrology and diabetology outpatient clinic decreased the odds of non-consent compared to the cardiology outpatient clinic.

Fig. 3
figure 3

Determinants associated with non-consent compared to consent, including 95% confidence intervals. METmin = metabolic equivalent of task-minutes; CVD = cardiovascular diseases; bpm = beats per minute; HbA1c = glycated haemoglobin; OR = odds ratio indicating the likelihood of having provided a non-consent as compared to a consent; % = percentage

Full size image

Exploratory analyses: determinants of non-response

The associations related to non-consent and non-response followed a similar pattern regarding age, heamoglobin, heartrate, BMI, and cardiovascular disease history (S7F1). Different was that a high education level, high physical activity level, peripheral arterial disease history, HbA1c level >  = 42 mmol/mol and low blood pressure were more strongly related to non-response compared to the associations found related to non-consent. Being referred to the geriatrics department and proteinuria, however, were not significantly associated with non-response.

Clinical implications of non-consent and non-response

Registration of extractable CVRM data from the EHR was 5 to 10% lower in non-consenting patients compared to consenting patients (Table 2). Moreover, availability reduced by more than 25% in items that were extracted from the UCC-CVRM questionnaire (history of CVD, smoking status, and hypertension history). Because of the impossibility of extracting such data form the EHR, 10-year risk estimates on cardiovascular morbidity or mortality were only available in 41% of non-consenting patients. The extractability of CVRM data in the non-response group was lower, ranging between the 2% and 50%.

Table 2 Extractability of the CVRM data in % in the EHR, stratified by consent status
Full size table

Discussion

Summary of the findings

We studied the differences between non-consenting, non-responding and consenting patients. Furthermore, we assessed whether cardiovascular risk indicator extractability from the EHR differed between the consent categories. We found that patients who were between 30–39 or 80–89 years old, less physically active, less educated, with a higher heartrate, higher HbA1c levels, lower haemoglobin levels, and a cardiovascular disease history were more often non-consenting. Non-response was associated with an age between 18–29 and 80–89, high education level, high physical activity level, low BMI, cardiovascular disease history, a high heart rate, low haemoglobin levels, low systolic blood pressure and higher HbA1c- levels.

Comparison with the literature

To our knowledge limited research has been conducted on the characteristics of non-consenting and non-responding patients, especially in cardiovascular LHSs targeting patients that are referred to the hospital for secondary prevention. Previous research on non-consent and, more often, non-response was mostly conducted in population-based research settings and clinical trials and not necessarily related to cardiovascular disease, severely limiting a good comparison between our results and the existing literature.

However, these studies found that non-consenting and non-responding patients generally were less educated [24,25,26,27], older [24, 28], women [28, 29], less physically active [30], more smokers [28], and had a higher disease burden [24, 28, 31, 32]. On the contrary, others reported non-responders to be younger [29, 31, 33].

Our findings regarding non-consent can partially be explained by the ‘worried well’ phenomenon: relatively healthy (younger) and educated patients tend to seek more medical advice and are more willing to participate in studies [26, 34, 35]. Participating in the UCC-CVRM LHS may give these patients the sense that their health is more closely monitored. An important addition from our study is that non-consent rates varied considerably across departments to which the patients were referred to for evaluation. We found that most of the non-consenting patients were referred to the geriatrics department and that they were older than the patients referred to the other departments. The elderly population is a heterogeneous group consisting of vital and frail elderly patients. Compared to more vital elderly patients, frail elderly patients are more at-risk for negative health outcomes [36]. The geriatrics department will be visited more by frail elderly patients, whereas other outpatient clinics will more likely be visited by more vital elderly patients. Previous research showed that frail elderly patients are more difficult to include into clinical studies because of the challenges associated with obtaining informed consent from this group [37, 38].

In addition, previous studies showed education level, among others, to be associated with health literacy [39, 40]. Others found that limited health literacy was associated with limited understanding of consent forms, leading to anxiety and less satisfaction with the consent process [41]. Clinical trial consent forms often appear to be written in an above average health literacy level, which, as a consequence, means that patients do not understand what they could be consenting for and, therefore, decline the invitation to participate [42]. Although the UCC-CVRM is not a clinical trial and our ethics committee demands writing patient information on a last class primary school level, this could be an explanation for the non-consenting patients more often being less educated.

The ‘worried-well’ phenomenon would, however, not entirely justify why patients with higher BMI and a history of proteinuria were at lower risk of non-consent. It is known that BMI changes with age and that the pattern of change is U-shaped. BMI increases in the younger age groups and then decreases in the older age groups [43, 44]. With the non-consent group being older, it is evident that our results point towards an association between lower BMI and non-consent. Nevertheless, generally healthy patients with an increased BMI could also be worried about their health and, therefore, agree to participate in a cardiovascular LHS. Data on proteinuria, however, was extracted from the self-reported UCC-CVRM questionnaire. The literal question was: “Have you ever been diagnosed with protein in your urine?” and, thus, not specifically stating increased protein levels in their urine, potentially leading patients with a normal amount of protein in their urine to answer this question with ‘yes’.

Surprising and in contrast to the associations found in the multivariable logistic regression analysis regarding the determinants of non-consent was the finding that a high education and physical activity level was associated with non-response. This might suggest that different profiles of patients are combined in the non-responding group: the patients with a condition already regularly controlled and the younger healthier group. These results should, however, be interpreted cautiously due to the significant missingness in the non-response group, potentially hampering the validity of these results. Nevertheless, and in line with previous research, non-responding patients seemed different from the consenting patients, but also from non-consenting patients [10, 27]. Indicating that, when studying non-participation, one should consider keeping the different categories of non-participation separate, even though non-response and non-consent are often considered to be the same.

Informed consent, participation bias on causal research

Our study indicates that an informed consent procedure in a cardiovascular LHS leads to a misrepresentation of the target population by the consenting patients, hampering generalisability of research results. We did, however, not assess whether this led to over -or underestimation of the associations between cardiovascular risk indicators in the target population compared to the associations within consenting patients. We recommend this for future research.

Informed consent and impact on learning healthcare systems

Routine clinical care data is increasingly used to improve the quality of care in a LHS approach as well as for research purposes. However, because of the apparent participation bias, the validity and generalisability of the results from research based on routine care data of the LHS could be at stake.

There is not yet a clear approach as to how patients should consent for the use of their routine clinical care data in a LHS design, as there is no consensus about the requirement of an informed consent [45]. However, a traditional informed consent procedure does not seem to be fit for purpose. The concerns about bias as a result of the traditional informed consent procedure in cohort studies have been expressed before [46]. Cumyn et al. [45] reviewed different types of consent forms, indicating meta-consent or dynamic consent as being the most appropriate within a LHS. Additionally, they emphasize the importance of information transfer between the professional and the patient. When communication about the LHS is lacking, patients are less likely to consent for the use of their data [45].

Impact on clinical care

Previous research from Groenhof et al. [47] compared the completeness of extractable CVRM indicators before and after the UCC-CVRM initiation in consenting patients. They showed that an infrastructure such as the UCC-CVRM LHS leads to a substantial improvement in the completeness of these indicators in the EHR, further enabling the use of, among others, cardiovascular risk algorithms, resulting in more information on the patient’s cardiovascular risk profile when determining the treatment strategy [47]. However, some patients in our study actively refused to participate in the LHS or did not respond to the invitation, which may have led to the less structured registration of CVRM indicators found in our study. This might affect the optimal CVRM in routine practice.

Conclusion

An informed consent procedure in a LHS may lead to participation bias. Furthermore, structured registration of CVRM indicators in the EHR was less in non-consenting and non-responding patients, which is detrimental for the LHS feedback loop, and potentially leads to suboptimal CVRM. This study underlines the importance of reassessing the need of a traditional informed consent procedure for the use of routine clinical care data in LHSs.

Availability of data and materials

The dataset analysed during the current study is not publicly available due to privacy and ethical concerns but is available from the corresponding author on reasonable request.

Abbreviations

LHS:

Learning Healthcare System

CVRM:

Cardiovascular riskmanagement

UMC:

University Medical Center

BMI:

Body Mass Index

UCC:

Utrecht Cardiovascular Cohort

EHR:

Electronic Health Record

LDL:

Low-Density Lipoprotein

MET:

Metabolic Equivalent of Task

SQUASH:

Short QUestionnaire To ASsess Health-enhancing Physical Activity

CVD:

Cardiovascular Disease

SCORE:

Systematic COronary Risk Evaluation

UKPDS:

United Kingdom Prospective Diabetes Study

SMART:

Second Manifestations of Arterial Disease

COVID:

Coronavirus Disease

ANOVA:

Analysis of Variance

MICE:

Multivariate Imputation by Chained Equations

OPD:

Outpatient Department

AMI:

Acute Myocardial Infarction

CAGB:

Coronary Artery Bypass Grafting

CHF:

Congestive Heart Failure

ICH:

Intracranial Hemorrhage

TIA:

Transient Ischemic Attack

PAD:

Peripheral Arterial Disease

AAA:

Abdominal Aortic Aneurysm

HDL:

High-Density Lipoprotein

Hb:

Haemoglobin

eGFR:

Estimated Glomerular Filtration Rate

CKD:

Chronic Kidney Disease

HbA1c:

Glycated haemoglobin

OR:

Odds Ratio

CI:

Confidence Interval

SBP:

Systolic blood pressure

DBP:

Diastolic blood pressure

References

  1. Jan S, Usherwood T, Brien JA, Peiris D, Rose J, Hayman N, et al. What determines adherence to treatment in cardiovascular disease prevention? Protocol for a mixed methods preference study. BMJ Open. 2011;1(2):1–7.

    Article  Google Scholar 

  2. Wiersma T, Smulders YM, Stehouwer CDA, Konings KTS, Lanphen J. Summary of the multidisciplinary guideline on cardiovascular risk management (revision 2011). Nederlands tijdschrift voor geneeskunde. 2012;156(36):A5104.

    PubMed  Google Scholar 

  3. WHO. HEARTS technical package for cardiovascular disease management in primary health care: risk based CVD management. Geneva: World Health Organization; 2020. p. 10–74.

    Google Scholar 

  4. Piepoli MF, Hoes AW, Agewall S, Albus C, Brotons C, Catapano AL, et al. 2016 European Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J. 2016;37(29):2315–81.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Kotseva K, Wood D, De Bacquer D, De Backer G, Rydén L, Jennings C, et al. EUROASPIRE IV: A European Society of Cardiology survey on the lifestyle, risk factor and therapeutic management of coronary patients from 24 European countries. Eur J Prev Cardiolog. 2016;23(6):636–48.

    Article  Google Scholar 

  6. Asselbergs FW, Visseren FLJ, Bots ML, De Borst GJ, Buijsrogge MP, Dieleman JM, et al. Uniform data collection in routine clinical practice in cardiovascular patients for optimal care, quality control and research: The Utrecht Cardiovascular Cohort. Eur J Prev Cardiol. 2017;24(8):840–7.

    Article  PubMed  Google Scholar 

  7. Groenhof TKJ, Lely AT, Haitjema S, Nathoe HM, Kortekaas MF, Asselbergs FW, et al. Evaluating a cardiovascular disease risk management care continuum within a learning healthcare system: a prospective cohort study. BJGP Open. 2020;4(5):1–10.

    Article  Google Scholar 

  8. Greene SM, Reid RJ, Larson EB. Implementing the learning health system: From concept to action. Ann Intern Med. 2012;157(3):207–10.

    Article  PubMed  Google Scholar 

  9. Groenhof TKJ, Mostert M, Lea NC, Haitjema S, de Vries MC, van Dijk WB, et al. How traditional informed consent impairs inclusivity in a learning healthcare system: lessons learned from the Utrecht Cardiovascular Cohort. J Clin Epidemiol. 2022. https://0-doi-org.brum.beds.ac.uk/10.1016/j.jclinepi.2022.04.025

  10. Ohlson CG, Ydreborg B. Participants and Non-participants of Different Categories in a Health Survey: A Cross-sectional Register Study. Scandinavian J Soc Med. 1985;13(2):67–74.

    Article  CAS  Google Scholar 

  11. Stuckless S, Parfrey PS. Bias in Clinical Research. In: Clinical Epidemiology: Practice and Methods [Internet]. 3rd ed. New York: Humana Press; 2016. https://0-doi-org.brum.beds.ac.uk/10.1007/978-1-0716-1138-8_2.

    Chapter  Google Scholar 

  12. Williams B, Irvine L, McGinnis AR, McMurdo MET, Crombie IK. When “no” might not quite mean “no”; the importance of informed and meaningful non-consent: Results from a survey of individuals refusing participation in a health-related research project. BMC Health Serv Res. 2007;7:59.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Galea S, Tracy M. Participation Rates in Epidemiologic Studies. Ann Epidemiol. 2007;17(9):643–53.

    Article  PubMed  Google Scholar 

  14. Keeble C, Baxter PD, Barber S, Law GR. Participation rates in epidemiology studies and surveys : A review 2007–2015. Inter J Epidemiol. 2015;14(1):1–14.

    Google Scholar 

  15. National Institute for Public Health and the Environment. First wave of COVID-19 had major impact on regular healthcare and health [Internet]. 2020 [cited 2022 Jul 26]. Available from: https://www.rivm.nl/en/news/first-wave-of-covid-19-had-major-impact-on-regular-healthcare-and-health

  16. Wendel-Vos GCW, Schuit AJ, Saris WHM, Kromhout D. Reproducibility and relative validity of the short questionnaire to assess health-enhancing physical activity. J Clin Epidemiol. 2003;56(12):1163–9.

    Article  PubMed  Google Scholar 

  17. Smulders YM, Burgers JS, Scheltens T, van Hout BA, Wiersma T, Simoons ML. Clinical practice guideline for cardiovascular risk management in the Netherlands. Neth J Med. 2008;66(4):169–74.

    CAS  PubMed  Google Scholar 

  18. Stevens RJ, Kothari V, Adler AI, Stration IM. The UKPDS risk engine: a model for the risk of coronary heart disease in Type II diabetes (UKPDS 56). Clin Sci. 2001;101(6):671.

    Article  CAS  Google Scholar 

  19. Dorresteijn JAN, Visseren FLJ, Wassink AMJ, Gondrie MJA, Steyerberg EW, Ridker PM, et al. Development and validation of a prediction rule for recurrent vascular events based on a cohort study of patients with arterial disease: the SMART risk score. Heart. 2013;99(12):866–72.

    Article  PubMed  Google Scholar 

  20. van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations in R. J Stat Soft [Internet]. 2011;45(3). [cited 2022 Aug 12]. Available from: http://www.jstatsoft.org/v45/i03/

  21. Kwak C, Clayton-Matthews A. Multinomial logistic regression. Nurs Res. 2002;51(6):404–10.

    Article  PubMed  Google Scholar 

  22. LaValley MP. Logistic regression. Circulation. 2008;117(18):2395–9.

    Article  PubMed  Google Scholar 

  23. R Development Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna: R Foundation for Statistical Computing; 2018. [cited 2022 Aug 31]. Available from: https://www.r-project.org/.

    Google Scholar 

  24. Chou P, Kuo HS, Chen CH, Lin HC. Characteristics of non-participants and reasons for non-participation in a population survey in Kin-Hu Kinmen. Eur J Epidemiol. 1997;13(2):195–200.

    Article  CAS  PubMed  Google Scholar 

  25. Christensen AI, Ekholm O, Gray L, Glümer C, Juel K. What is wrong with non-respondents? Alcohol-, drug- and smoking-related mortality and morbidity in a 12-year follow-up study of respondents and non-respondents in the Danish Health and Morbidity Survey. Addiction. 2015;110(9):1507–14.

    Article  Google Scholar 

  26. Boshuizen HC, Viet AL, Picavet HSJ, Botterweck A, van Loon AJM. Non-response in a survey of cardiovascular risk factors in the Dutch population: Determinants and resulting biases. Public Health. 2006;120(4):297–308.

    Article  CAS  PubMed  Google Scholar 

  27. Harder E, Juul KE, Jensen SM, Thomsen LT, Frederiksen K, Kjaer SK. Factors associated with non-participation in cervical cancer screening – A nationwide study of nearly half a million women in Denmark. Prev Med. 2018;111:94–100.

    Article  PubMed  Google Scholar 

  28. Enzenbach C, Wicklein B, Wirkner K, Loeffler M. Evaluating selection bias in a population-based cohort study with low baseline participation: The LIFE-Adult-Study. BMC Med Res Methodol. 2019;19(1):1–14.

    Article  Google Scholar 

  29. Batty GD, Gale CR. Impact of resurvey non-response on the associations between baseline risk factors and cardiovascular disease mortality: Prospective cohort study. J Epidemiol Commun Health. 2009;63(11):952–5.

    Article  CAS  Google Scholar 

  30. Hoebel J, Starker A, Jordan S, Richter M, Lampert T. Determinants of health check attendance in adults: Findings from the cross-sectional German Health Update (GEDA) study. BMC Public Health. 2014;14(1):1–12.

    Article  Google Scholar 

  31. Gijsberts CM, Den Ruijter HM. Non-response to questionnaires independently predicts mortality of coronary angiography patients. Int J Cardiol. 2015;201:168–70.

    Article  PubMed  Google Scholar 

  32. Hassani M, Kivimaki M, Elbaz A, Shipley M, Singh-Manoux A, Sabia S. Non-consent to a wrist-worn accelerometer in older adults: the role of socio-demographic, behavioural and health factors. Reddy H, editor. PLoS ONE. 2014;9(10):e110816.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Cochrane T, Gidlow CJ, Kumar J, Mawby Y, Iqbal Z, Chambers RM. Cross-sectional review of the response and treatment uptake from the NHS Health Checks programme in Stoke on Trent. J Public Health (United Kingdom). 2013;35(1):92–8.

    Google Scholar 

  34. Chatterjee SS, Vora M, Malathesh BC, Bhattacharyya R. Worried well and Covid-19: Re-emergence of an old quandary. Asian J Psychiatry. 2020;54:102247.

    Article  Google Scholar 

  35. Badenbroek IF, Nielen MMJ, Hollander M, Stol DM, Drijkoningen AE, Kraaijenhagen RA, et al. Mapping non-response in a prevention program for cardiometabolic diseases in primary care: How to improve participation? Prev Med Rep. 2020;2020(19):101092.

    Article  Google Scholar 

  36. Kojima G, Liljas A, Iliffe S. Frailty syndrome: implications and challenges for health care policy. RMHP. 2019;12:23–30.

    Article  Google Scholar 

  37. Hempenius L, Slaets JPJ, Boelens MAM, van Asselt DZB, de Bock GH, Wiggers T, et al. Inclusion of frail elderly patients in clinical trials: Solutions to the problems. J Geriatr Oncol. 2013;4(1):26–31.

    Article  PubMed  Google Scholar 

  38. Barron JS, Duffey PL, Byrd LJ, Campbell R, Ferrucci L. Informed consent for research participation in frail older persons. Aging Clin Exp Res. 2004;16(1):79–85.

    Article  PubMed  Google Scholar 

  39. Bayati T, Dehghan A, Bonyadi F, Bazrafkan L. Investigating the effect of education on health literacy and its relation to health-promoting behaviors in health center. J Edu Health Promot. 2018;7(1):127.

    Article  Google Scholar 

  40. van der Heide I, Wang J, Droomers M, Spreeuwenberg P, Rademakers J, Uiters E. The relationship between health, education, and health literacy: results from the Dutch adult literacy and life skills survey. J Health Commun. 2013;18(1):172–84.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Burks AC, Keim-Malpass J. Health literacy and informed consent for clinical trials: a systematic review and implications for nurses. Nurs Res Rev. 2019;9:31–40.

    Article  Google Scholar 

  42. Simonds VW, Garroutte EM, Buchwald D. Health literacy and informed consent materials: designed for documentation, not comprehension of health research. Physiol Behavior. 2017;176(12):139–48.

    Google Scholar 

  43. Drøyvold WB, Nilsen TIL, Krüger O, Holmen TL, Krokstad S, Midthjell K, et al. Change in height, weight and body mass index: Longitudinal data from the HUNT Study in Norway. Int J Obes. 2006;30(6):935–9.

    Article  Google Scholar 

  44. Reas DL, Nygård JF, Svensson E, Sørensen T, Sandanger I. Changes in body mass index by age, gender, and socio-economic status among a cohort of Norwegian men and women (1990–2001). BMC Public Health. 2007;7:1–7.

    Article  Google Scholar 

  45. Cumyn A, Barton A, Dault R, Cloutier AM, Jalbert R, Ethier JF. Informed consent within a learning health system: A scoping review. Learn Health Syst. 2020;4(2):1–14.

    Google Scholar 

  46. Kho ME, Duffett M, Willison DJ, Cook DJ, Brouwers MC. Written informed consent and selection bias in observational studies using medical records: Systematic review. BMJ. 2009;338(7698):822.

    Google Scholar 

  47. Groenhof TKJ, Haitjema S, Lely AT, Grobbee DE, Asselbergs FW, Bots ML. Optimizing cardiovascular risk assessment and registration in a developing cardiovascular learning healthcare system: women benefit most. In: Towards a Learning Healthcare System. Enschede: Gildeprint B.V; 2020. p. 23–39.

    Google Scholar 

Download references

Acknowledgements

Members of the Utrecht Cardiovascular Cohort- CardioVascular Risk Management (UCC-CVRM) Study group: F.W. Asselbergs, Department of Cardiology; G.J. de Borst, Department of Vascular Surgery; M.L. Bots (chair), Julius Center for Health Sciences and Primary Care; M. Hollander, Julius Center for Health Sciences and Primary Care; M.H. Emmelot, Department of Geriatrics; P.A. de Jong, Department of Radiology; A.T. Lely, Department of Obstetrics/Gynecology; I.E. Hoefer, Central Diagnostic Laboratory; N.P. van der Kaaij, Department of Cardiothoracic Surgery; Y.M. Ruigrok, Department of Neurology; M.C. Verhaar, Department of Nephrology & Hypertension, F.L.J. Visseren, Department of Vascular Medicine, University Medical Center Utrecht and Utrecht University.

UCC-CVRM study group

Pim A. de Jong4, Marianne C. Verhaar5, Frank L.J. Visseren6, Folkert W. Asselbergs7, Niels P. van der Kaaij8, Imo E. Hoefer1, Gert J. de Borst9, Ynte M. Ruigrok10, Monika Hollander3, A. Titia Lely11, Mariëlle H. Emmelot-Vonk12, Michiel L. Bots3.

4Department of Radiology, University Medical Center Utrecht, Utrecht, The Netherlands

5Department of Nephrology & Hypertension, University Medical Centre Utrecht, Utrecht, the Netherlands

6Department of Vascular Medicine, University Medical Center Utrecht, Utrecht, The Netherlands

7Department of Cardiology, University Medical Center Utrecht, Utrecht University, Utrecht, the Netherlands

8Department of Cardiothoracic Surgery, University Medical Center Utrecht, The Netherlands

9Department of Vascular Surgery, University Medical Center Utrecht, Utrecht Netherlands

10UMC Utrecht Brain Center, Department of Neurology and Neurosurgery, University Medical Center Utrecht, University Utrecht, Utrecht, The Netherlands

11Department of Obstetrics, Wilhelmina Children’s Hospital Birth Center, University Medical Center Utrecht, The Netherlands

12Department of Geriatric Medicine, University Medical Centre Utrecht, Utrecht, Netherlands

Funding

The UCC-CVRM is primarily financed by the UMC Utrecht. A grant from the Netherlands Organization for Health Research and Development (#8480–34001) was obtained to develop feedback procedures. UCC-CVRM website: www.umuctrecht.nl/ucc (in Dutch). Contact information UCC-CVRM: ucc@umcutrecht.nl. Anna Zondag was supported by a grant from the European Union’s Horizon 2020 research and innovation program under grant agreement No. 101017331 (ODIN).

The funding sources were not involved in the study design, analysis, and interpretation of the data, nor in the writing of the manuscript and in the decision to submit the manuscript for publication.

Author information

Authors and Affiliations

  1. Central Diagnostic Laboratory, Division of Laboratories, Pharmacy and Biomedical Genetics, University Medical Center Utrecht, Utrecht University, Utrecht, the Netherlands

    Anna G. M. Zondag, Wouter W. van Solinge & Saskia Haitjema

  2. Department of Obstetrics and Gynaecology, St. Antonius Hospital, Utrecht, the Netherlands

    T. Katrien J. Groenhof

  3. Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands

    T. Katrien J. Groenhof, Rieke van der Graaf & Michiel L. Bots

Authors
  1. Anna G. M. Zondag
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. T. Katrien J. Groenhof
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Rieke van der Graaf
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Wouter W. van Solinge
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Michiel L. Bots
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Saskia Haitjema
    View author publications

    You can also search for this author in PubMed Google Scholar

Consortia

UCC-CVRM study group

  • Pim A. de Jong
  • , Marianne C. Verhaar
  • , Frank L. J. Visseren
  • , Folkert W. Asselbergs
  • , Niels P. van der Kaaij
  • , Imo E. Hoefer
  • , Gert J. de Borst
  • , Ynte M. Ruigrok
  • , Monika Hollander
  • , A. Titia Lely
  • , Mariëlle H. Emmelot-Vonk
  •  & Michiel L. Bots

Contributions

A.G.M.Z., W.W.v.S., M.L.B and S.H. contributed to the conceptualization of the project and methodology. W.W.v.S., M.L.B and S.H. contributed to the financial acquisition. A.G.M.Z., T.K.J.G., W.W.v.S., M.L.B, S.H. and R.v.d.G contributed to the analysis and interpretation of the data and critically reviewed the manuscript. A.G.M.Z. wrote the draft manuscript. Approval of the final manuscript was obtained by all authors.

Corresponding author

Correspondence to Anna G. M. Zondag.

Ethics declarations

Ethics approval and consent to participate

The UCC-CVRM was approved by the Institutional Review Board (IRB) of the UMC Utrecht, NedMec [6]. For this current study into the differences between consenting, non-consenting and non-responding patients, informed consent was waived by the IRB of the UMC Utrecht (number 19/641). Also, all methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1:

Supplement 1. STROBE checklist for reporting. Supplement 2. UCC-CVRM questionnaire. Supplement 3. Missingness. Supplement 4. Patient inclusion flow-chart. Supplement 5. Sensitivity analysis to explore age distributions. Supplement 6. Determinants of non-consent. Supplement 7. Determinants of non-response (exploratory analysis).

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zondag, A.G.M., Groenhof, T.K.J., van der Graaf, R. et al. Asking informed consent may lead to significant participation bias and suboptimal cardiovascular risk management in learning healthcare systems. BMC Med Res Methodol 23, 98 (2023). https://0-doi-org.brum.beds.ac.uk/10.1186/s12874-023-01924-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://0-doi-org.brum.beds.ac.uk/10.1186/s12874-023-01924-6

Keywords

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us