Introduction

Breaking bad news to a patient or a next of kin is a challenging conversation as the information disclosed often has an altering effect on the person’s life perspective. Herein, the way how bad news are communicated might play an important role for patients' psychological burden.1,2,3,4,5 Several communication techniques and guidelines such as specific communication protocols6 were developed, to facilitate the disclosure of bad news to patients and relatives.

For a long time, experts recommended disclosing bad news in person whenever possible2, 3, 7 as it renders addressing patients’ or relatives’ emotional responses more easily. Still, it was acknowledged that in certain situations it is more feasible to disclose bad news by telephone.8 Particularly, in case of a clinical deterioration or even sudden death of a patient, the imminent disclosure of the bad news over the phone might spare the next of kin a prolonged time of fearful uncertainty.9, 10 Around the year 2000, approximately one quarter of patient-physician conversations were conducted via telephone11 and with further development of mobile communication technologies and a growing need for cost-effective treatments, telephone consultations have become even more common. In fact, the current COVID-19 pandemic brought a sudden increase in telemedicine in order to minimize the risk of spreading the virus12, 13 and due to hospitals’ visitation restrictions. Since 2020, medical conversations via telephone including the disclosure of bad news are often a necessary substitute for in-person appointments and became an integral part of clinical practice across the world.

Therefore, further insight regarding the psychological impact of breaking bad news by telephone on patients and next of kin compared to breaking bad news in person is needed.

The aim of this systematic review and meta-analysis was to investigate whether disclosure of bad news by telephone is an appropriate alternative to in-person disclosure in terms of psychological distress and satisfaction with care measured by symptoms of anxiety, depression, and post-traumatic stress disorder (PTSD), as well as patient satisfaction.

Methods

Types of Studies, Participants, and Outcomes

We conducted this systematic review and meta-analysis in accordance with the updated version of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting guidelines (PRISMA 2020)14 and registered it in the International Prospective Register of Systematic Reviews (PROSPERO; ID: PROSPERO 2021 CRD42021233266). We included peer-reviewed observational studies, randomized controlled trials (RCT) and quasi-RCTs that investigated differences in psychological distress of breaking bad news by telephone compared to in person in patients or next of kin.

Our primary outcome was psychological distress defined as symptoms of anxiety, depression or PTSD. Our secondary outcomes were satisfaction with care including trust in the healthcare worker disclosing the bad news. Studies were eligible if they reported results on at least one of our primary or secondary outcomes. No restrictions concerning age or gender of adult participants and no publication date restrictions were applied.

Exclusion criteria were 1) no participants ≥ 18 years, 2) studies only including patients with psychiatric diagnoses, 3) studies only including patients with moderate to severe cognitive impairment, 4) no comparison of telephone versus in-person disclosure, 5) no results on at least one of the primary and secondary outcomes, and 6) conference articles or abstracts and case reports.

This manuscript is based on the MOOSE Checklist of Meta-analyses and Observational Studies.15

Database Search for the Identification of Studies

We searched the digital databases PubMed, Embase, CINAHL and PsycInfo using a string of search terms consisting of subject headings and free-text words which we had developed together with an academic librarian experienced in systematic reviews (C.A.-H.). The search strings and filters for each database search can be found in the Supplementary Material. To identify additional studies, we screened all references of eligible studies through the cited references search of Web of Science and PubMed. The latest search was performed on October 18, 2022.

Study Selection

Two investigators screened the titles and abstracts of articles regarding inclusion and exclusion criteria and independently assessed the full texts of all remaining studies. Disagreements were resolved through discussion with a third reviewer. Two investigators independently extracted the relevant data from the included studies.

Risk of Bias Evaluation

We evaluated the risk of bias for every relevant outcome of all included studies using The Cochrane Collaboration’s tool for assessing risk of bias.16 Two authors independently assessed the risk of bias for all studies and resolved disagreements by discussion until consensus was found. A detailed description of the risk of bias assessment can be found in the Supplementary Material.

Analysis

We synthesized the findings on primary and secondary outcomes of all studies in a qualitative analysis. Studies that provided data on the mean and standard deviation of psychometric scores assessing one of the outcomes and/or the numbers of patients with and without the outcome were included in the meta-analysis. We pooled continuous data using random-effects models and reported the standardized mean difference, i.e. inverse variance (IV) with 95% confidence intervals (CI). Heterogeneity was examined through visual inspection of the forest plots. We evaluated dichotomous data with a random-effects model applying the Mantel–Haenszel method and reported odds ratios (OR) and 95% confidence intervals. For the latter, we used the I2 statistic, which quantifies inconsistency across studies, to assess the consequences of heterogeneity on the meta-analysis. For all analyses, a two-sided p-value < 0.05 was considered statistically significant. Statistical analyses were conducted using the METAN package in Stata (Stata MP, version 15.1; StataCorp LP).

Results

Study Selection

We identified 5944 records through the database search and three through citation tracking. After removing 1514 duplicates, we screened 4433 records based on titles and abstracts and in the process excluded 4216. Two reviewers independently reviewed the full texts of 214 articles and were thus able to include nine. There was one additional eligible article17 that we did not include as it reported a secondary analysis of a study we had already included,18 with the same analyses and outcome parameters, only at different follow-up time points. Citation tracking yielded two further records eligible for inclusion which led to 11 studies being included in the qualitative synthesis and 9 in the meta-analysis (Fig. 1).

Figure 1
figure 1

PRISMA flow diagram of the study selection process.

Full size image

Description of Studies

The 11 included studies were published between 1997 and 2021 and predominantly conducted in the USA10, 18,19,20,21,22 as well as in Australia23 and European countries, i.e., the United Kingdom,24 the Netherlands,25 Germany,15 and Denmark.26 Study sample sizes ranged from 24 to 434 participants. Seven studies evaluated the disclosure of malignancy diagnoses such as breast cancer,15, 22, 24 gynecologic cancer [19], thoracic cancer,26 melanoma,23 and different types of cancer.10 The remaining 4 studies assessed breaking bad news of genetic testing results for high-risk constellations of hereditary breast, gynecological and/or gastrointestinal cancer,18, 21 Alzheimer disease20 and hypertrophic cardiomyopathy.25

All studies investigating psychological distress used well-established and validated questionnaires. Assessment of satisfaction varied broadly from singular binary or categorical items to scales and qualitative interviews. Follow-up times ranged from several days to years.

The study characteristics and main findings are presented in Table 1.

Table 1 Summary of studies included in the systematic review
Full size table

Description of Findings of the Included Studies

Association of Disclosure of Bad News via Telephone vs. in Person with Psychological Distress

Anxiety Symptoms

Five out of all 11 studies published in 2003 and 2016 to 2021 reported findings on the association of disclosure of bad news via telephone vs. in person with symptoms of anxiety at follow-up.18, 20, 21, 23, 26 None of the studies showed increased anxiety in patients to whom bad news were disclosed via telephone compared to those with in-person disclosure.

Three randomized controlled trials investigated the disclosure of a positive genetic test result. In the study of Bradbury et al.,18 patients with telephone disclosure showed a greater decrease in general anxiety, but not state anxiety, i.e. a transient emotional state, from baseline (pre-disclosure) to one week post-disclosure compared to those with in-person disclosure when raw data rather than the imputed data set was analyzed. In the study of Christensen et al.,20 telephone disclosure was non-inferior to in-person disclosure regarding anxiety symptoms at 6-week, 6-month and 12-month follow-up. In the study of Kinney et al.,21 anxiety symptom levels one year after baseline did not differ between patients with telephone vs. in-person disclosure.

Two studies, a randomized controlled trial26 and a prospective cohort study23 evaluated the disclosure of malignancy diagnoses, i.e. lung, mediastinal & pleural cancers and melanoma, respectively. Both did not reveal any differences between patients with telephone and in-person disclosure regarding anxiety symptom levels 4 weeks and 4, 8 and 17 months after baseline, respectively.

Depressive Symptoms

Four out of 11 studies published in 2003 and 2016 to 2021 reported findings on the association of disclosure of bad news via telephone vs. in person with depressive symptoms at follow-up.18, 20, 23, 26 None found an association between telephone or in-person disclosure and depressive symptoms.

Two randomized controlled studies investigated disclosure of a positive genetic test result. In the study of Christensen et al.,20 telephone disclosure was inferior to in-person disclosure regarding depressive symptom levels at 12-month follow-up. There was no difference in depressive symptom levels between the two groups at 6-week and 6-month follow-up. Additionally, the average depressive symptom score was still well below the cutoff for clinical concern. In the study of Bradbury et al.,18 there was no association between telephone or in-person disclosure and change in depressive symptom levels from baseline to one week post-disclosure.

Two studies, a randomized controlled trial26 and a prospective cohort study23 evaluated disclosure of malignancy diagnoses, i.e. lung, mediastinal and pleural cancers and melanoma, respectively. Both did not reveal any differences in depressive symptom levels at 4 weeks and 4, 8 and 17 months after baseline, respectively.

PTSD Symptoms

Three studies published in 2003, 2018 and 2021 reported findings on the association of disclosure of bad news via telephone or in person and PTSD symptom levels at follow-up with one study revealing an association.18, 20, 21 All 3 studies are randomized controlled trials in the field of genetic testing. In the study of Christensen et al.,20 telephone disclosure was inferior to in-person disclosure regarding PTSD symptom levels at 12-month follow-up. There was no difference in symptom levels between the two groups at 6-week and 6-month follow-up. Bradbury et al.18 found that the change in PTSD symptom levels from baseline to one-week follow-up did not differ between patients with telephone vs. in-person disclosure. In the study of Kinney et al.,21 there was no change in PTSD symptom levels from baseline to follow-up within both groups. PTSD symptom levels 12 months after baseline did not differ between patients with telephone vs. in-person disclosure.

Association of Disclosure of Bad News via Telephone vs. in Person with Patient Satisfaction

Eight studies published in 1997 to 2019 reported findings on the association of disclosure of bad news via telephone vs. in person and patient satisfaction at follow-up with inconclusive results.10, 15, 18, 19, 22,23,24,25 In 3 of these, telephone disclosure was associated with lower satisfaction and in 2 with higher satisfaction. Lastly, 3 studies did not show any association between telephone or in-person disclosure and satisfaction.

Two studies evaluated disclosure of a positive genetic test result. In the randomized controlled trial of Bradbury et al.,18 there was no difference in patient satisfaction one week after disclosure between patients who received results via telephone vs. in person. In the cross-sectional study of Christiaans et al.,25 disclosure by telephone or mail was associated with higher patient satisfaction at 3-year follow-up compared to in-person disclosure.

Six observational studies evaluated satisfaction with disclosure of a new cancer diagnosis via telephone vs. in person. Three studies found that patients who received the diagnosis via telephone were less satisfied with disclosure compared to those who were told in person.10, 19, 22 Two studies did not reveal an association between telephone or in-person disclosure and patient satisfaction within the subsequent 6 weeks as well as 4, 8 and 17 months, respectively.15, 23 In the study of Campbell et al.,24 patients who received the bad news via telephone were more likely to be satisfied to have been informed this way than patients who were told in person.

Trust in the Health Care Worker Disclosing the Bad News

The study of Figg et al.10 evaluated the association of disclosure of bad news via telephone vs. in person and patients’ trust in physician after result disclosure. Almost 80% of patients in this sample reported a greater than neutral level of trust and 16% said they had absolute trust. There was no association between level of trust and disclosure of bad news via telephone vs. in person.

Quantitative Analysis

Nine studies with 1284 patients that evaluated breaking bad news via telephone compared to in person were included in the meta-analysis. Three studies reported results on psychological distress, i.e., anxiety, depression or PTSD, and 7 on satisfaction.

Anxiety Symptoms

Three studies (published between 2018 and 2021) including 285 patients evaluated symptoms of anxiety, 2 with a high risk of recruitment bias18, 26 and one with a low risk of bias.20 There was no mean difference regarding anxiety symptom levels when bad news was disclosed by telephone compared to in person (standardized mean difference [SMD] 0.10 [95% CI -0.15 to 0.35]) (Table 2). There was little heterogeneity among trials (I2 = 13%, p = 0.32).

Table 2 Association between disclosure of bad news via telephone vs. in person and symptoms of anxiety
Full size table

Depressive Symptoms

Three studies (published between 2018 and 2021) including 284 patients evaluated depressive symptoms, 2 with a high risk of recruitment bias18, 26 and one with a low risk of bias.20 There was no mean difference in depressive symptom levels when bad news was disclosed by telephone compared to in person (SMD 0.10 [95% CI -0.30 to 0.49]) (Table 3). There was substantial heterogeneity among trials (I2 = 64%, p = 0.06).

Table 3 Association between disclosure of bad news via telephone vs. in person and depressive symptoms
Full size table

PTSD Symptoms

Two studies assessed symptoms of PTSD in 171 patients with a high risk of recruitment bias18 and low risk of bias,20 respectively. There was no mean difference in symptom levels of PTSD when bad news was disclosed by telephone compared to in person (SMD -0.01 [95% CI -0.48 to 0.36]) (Table 4). Heterogeneity between trials was low (I2 = 0%, p = 0.74).

Table 4 Association between disclosure of bad news via telephone vs. in person and symptoms of post-traumatic stress disorder
Full size table

Satisfaction

Seven studies (published between 2009 and 2019) with mostly high risk of bias evaluated satisfaction with four studies including 678 patients assessing satisfaction levels10, 18, 19, 25 and 3 studies with 409 participants comparing the proportions of patients who were satisfied with the way bad news were disclosed.22,23,24 There was no mean difference in satisfaction levels when bad news were disclosed by telephone compared to in person (SMD -0.29 [95% CI -0.83 to 0.25]) (Table 5). Further, risk for low satisfaction in patients who received bad news by telephone was similar compared to those with in-person disclosure (OR 1.00 [95% CI 0.26 to 3.84]) (Fig. 2). Heterogeneity among these trials was high (I2 = 87%, p = 0.0005).

Table 5 Association between the disclosure of bad news via telephone vs. in person and satisfaction levels
Full size table
Figure 2
figure 2

Forest plot showing the association between disclosure of bad news via telephone vs. in person and patient satisfaction. Legend: The squares and horizontal lines correspond to the study-specific odds ratios (OR) and 95% confidence intervals (CI), respectively. The diamond represents the pooled OR of satisfaction. Abbreviations: M-H = Mantel–Haenszel method.

Full size image

Discussion

There is a growing demand for telemedicine including the disclosure of bad news despite little insight regarding potential adverse effects. Therefore, in this systematic review and meta-analysis, we investigated if disclosure of bad news via telephone is associated with increased psychological distress and lower patient satisfaction compared to in-person settings. We included 11 studies in the qualitative synthesis and 9 in the meta-analysis. Our findings suggest that breaking bad news via telephone is neither associated with increased psychological distress nor lower patient satisfaction compared to breaking bad news in person.

Five studies18, 20, 21, 23, 26 evaluated the association between the disclosure of bad news via telephone compared to in-person disclosure and psychological distress. None of the studies revealed any significant association between mode of disclosure and psychological distress, i.e., symptoms of anxiety, depression or PTSD. These studies evaluated the disclosure of a new cancer diagnoses and genetic test result indicating a high risk for Alzheimer disease or a hereditary cancer syndrome. Although, the types of bad news might differ regarding their immediate impact on patient’s life, results were similar.

The uniformity of results within and across the studies assessing different contents of bad news and manifestations of psychological distress at different time points, suggests that breaking bad news via telephone is not associated with increased psychological distress and could be an acceptable alternative to in-person disclosure, at least in certain settings. Indeed, when in-person disclosure of bad news within a reasonable time period is not possible, e.g. if a patient lives far away or there is no available appointment in the near future, and at the same time, receipt of the bad news is either urgent or a delay could trouble patients due to uncertainty, telephone disclosure may be preferable over in-person disclosure.

So far, there are few expert recommendations on breaking bad news via telephone to patients27,28,29 but these are mostly based on clinical experience and studies on how to communicate bad news in general and there are no specific evidence-based recommendations.5 So far, the findings of the existing studies suggest that the modality of disclosure might play a secondary role and the way in which the bad news are communicated might be more important. This might include preparing patients for the possibility of receiving bad news beforehand and, at the time of the conversation, first ensuring that they are in an appropriate setting.13, 28 Further, the structure and content of the breaking bad news conversation may be relevant. Therefore, several communication strategies were developed.6 In all of the five studies evaluating psychological distress that we included, the bad news were disclosed by specifically trained staff, i.e. genetic counselors and physicians who had completed courses in patient communication during their specialist training.

Due to restrictions during the coronavirus pandemic, doctor-patient consultations via video-chat have become more common. In comparison to the disclosure of bad news over the phone, videoconferencing offers the opportunity for doctors to identify non-verbal communication and through this facilitate the recognition of patients’ emotional concerns.

Recently, proposed adaptations of existing breaking bad news communication strategies for telephone and videoconference disclosure, which were based on clinical experience and experimental pilot studies, have been published.5, 13, 28,29,30 These adaptations include recommendations on ensuring that patients are in an appropriate setting which may involve their significant others, exploring and acknowledging emotions verbally and expressing empathy through tone of voice. These need to be complemented by further research on the topic to facilitate the development of evidence-based communication strategies and should also include virtual patient encounters.

While satisfaction among individual trials showed both, positive and negative associations with bad news disclosed over the phone compared to in-person disclosure, in the quantitative analysis there was no statistically significant difference. This may be explained by patient preferences regarding mode of disclosure. Yet, most of the included studies did not report patients’ preferences. Of note, two studies evaluating disclosure of genetic test results 18, 26 reported that a significant number of patients declined participation due to a preference for one of the two disclosure modes. In the study of Bradbury et al., 18 conducted in the United States, almost 20% of patients declined participation due to a preference for in-person disclosure. The study of Bodtger et al.26 which was conducted in Denmark, reported that 151 (31%) patients did not agree to randomization and 105 (70%) of those chose to receive their genetic test result via telephone. Patients’ preference regarding the mode of disclosure and the involvement in decision-making31 might be associated with their satisfaction with the disclosure and should be evaluated in further research. Further, several studies revealed that other factors such as length of the conversation and discussion of treatment options were associated with patient satisfaction.10, 19 One study assessed patients’ level of trust.10 This study reports that almost 80% of patients had greater than neutral trust in their clinician and longer conversations and the discussion of treatment options were associated with high trust. Patients’ trust in their physician following a disclosure of bad news might be less depent on the mode of disclosure but rather on the quality of the relationship and the way in which the bad news are communicated, e.g. with the physician showing empathy and offering support. This might further emphasize the importance of the quality of the breaking bad news conversation irrespective from mode of disclosure.

There is a wide range of publication years across the studies included for the evaluation of the secondary endpoint, i.e. patient satisfaction. Importantly, telemedicine has recently become more popular and patients may thus be more used to telephone consultations today compared to some time ago. This may impact the generalizability of results to today’s standard of care. The wide range of publication years across the studies on the association between mode of disclosure and patient satisfaction as well as the changing role of telephone consultations is an important point that needs to be considered. As telephone consultations even for difficult conversations have become much more common in many countries since, patients might perceive this as usual care, potentially impacting patient satisfaction.

Limitations

The 11 studies we were able to include, were heterogeneous regarding study design, patient populations, content of bad news, e.g. cancer diagnosis and increased genetic risk for a certain disease, and follow-up durations. This was especially true for the studies on our secondary outcome, i.e. satisfaction, which were also published over a time span of over 20 years. Due to this heterogeneity, generalization of our findings on the association between mode of disclosure and patient satisfaction is limited and further research is needed to confirm our findings. However, regarding psychological distress, 4 out of 5 studies included in the meta-analyses were RCTs with a methodologically sound study design published very recently between 2018 and 2021.

Three studies evaluated the disclosure of a high genetic risk for a certain disease or illness and two studies assessed the disclosure of a new malignancy diagnosis. Bad news are considered information that can potentially influence a patient’s life in some negative or unfavorable way. Still, the impact on patients’ psychological distress and satisfaction might differ. Due to the small number of identified studies, we were not able to analyze these two groups of studies separately.

Further, the use of telephone consultations including the disclosure of bad news has changed significantly since the beginning of the COVID-19 pandemic. While it was an option for patients who lived in great distance or preferred to be informed via telephone, during the COVID-19 pandemic suddenly it was often the only possible option. As all studies included in our systematic review were conducted before the beginning of the COVID-19 pandemic, this might limit the transferability of our findings to current times.

None of the studies calculated multivariable models including relevant covariates. As telephone disclosure of bad news is more common in the USA due to often great geographical distance between patient and treating healthcare worker, it might less likely lead to increased psychological distress and lower satisfaction. Our findings did not support this hypothesis. However, due to the small number of included studies, it is not possible to draw further conclusions.

Strengths

Based on an extensive literature search, this systematic review and meta-analysis presents the current state of research on relevant patient-related outcomes associated with breaking bad news via telephone compared to in person.

The disclosure of bad news is a relevant part of clinical practice and one of the most publicized topics in the field of communication in healthcare. According to Pubmed, more than 2000 articles regarding the disclosure of bad news have been published since the year 2000. However, empirical studies are still scarce and recommendations on the communication of bad news are mainly based on expert opinions and clinical experience. Interestingly, we only found 11 studies and only 4 RCTs investigating the effect of face-to-face compared to over the phone disclosure of bad news.

It is worth mentioning that we did not find any study on the disclosure of bad news to adult patients’ relatives in our systematic review. Research has shown that the way healthcare professionals communicate with relatives of patients that are dying may influence their long-term psychological well-being.32 Breaking bad news to relatives frequently occurs in case of severe medical conditions such as an accident, acute deterioration or death of the patient. In these situations, the disclosure of the bad news is usually more time-sensitive and more likely to be conducted over the phone. Hence, rigorous studies on how to disclose bad news with patients and relatives in person, over the phone or virtually are warranted.

Summary and Conclusions

Our findings suggest that disclosure of bad news via telephone compared to in-person disclosure does not lead to increased short- or long-term psychological distress, i.e. symptoms of anxiety, depression and PTSD or to lower satisfaction.

Since the beginning of the Covid-19 pandemic, there was an important increase of telephone disclosure of bad news as it was often the only available option when in-person consultations were not possible due to restrictions related to the risk of infection. Our results suggest that disclosure of bad news via telephone might be acceptable for patients and might not have adverse effects if the disclosure is well-conducted. Further insight on the association between the disclosure of bad news via telephone vs. in person as well as the role of other factors is needed to facilitate evidence-based recommendations and guidance for these challenging conversations.