“1 in 2 cancer patients significantly distressed”? Fake news from Psycho-Oncology

Study in Wiley journal Psycho-Oncology generates fake facts to promote cancer as a mental health problem and portray cancer patients in need of mandatory screening for distress.

The  exaggerated rates of significant distress stem from poor measurement and biased opportunistic sampling.

Results of this large study will not generalize to other samples using the same measures.

Publication of this study in the official journal of  the International Psycho-Onocology Society (IPOS)  reflects the organization’s  explicit political and professional goals-

  • to promote distress as the  6th vital sign.
  • to mandate screening for distress in all medical settings offering cancer care.

A systematic review has shown Screening for distress has not been shown to improve patient outcomes

Imposing mandated routine screening in well functioning cancer care settings may disrupt the delivery of informal support integrated to that care, curtail care to patients with more intensive needs, and increase inappropriate prescription of psychotropic drugs.

edited distress thermometerMuch better studies with validated instruments replicate that  1/3 of cancer patients in waiting rooms are distressed, approximately the same as rates in primary care waiting rooms. For instance, our study of breast cancer patients.

Authors of this study are not early career investigators needing a publication, but senior leaders of the field in Germany with funding and a registered study protocol.

The study

Mehnert A, Hartung TJ, Friedrich M, Vehling S, Brähler E, Härter M, Keller M, Schulz H, Wegscheider K, Weis J, Koch U. One in two cancer patients is significantly distressed: prevalence and indicators of distress. Psycho‐Oncology. 2017 May 31.


Objective: Psychological distress is common in cancer patients, and awareness of its indicators is essential. We aimed to assess the prevalence of psychological distress and to identify problems indicative of high distress.

Methods: We used the distress thermometer (DT) and its 34‐item problem list to measure psychological distress in 3724 cancer patients (mean age 58 years; 57% women) across major tumor entities, enrolled in an epidemiological multicenter study. To identify distress‐related problems, we conducted monothetic analyses.

Results: We found high levels of psychological distress (DT ≥ 5) in 52% of patients. The most prevalent problems were fatigue (56%), sleep problems (51%), and problems getting around (47%). Sadness, fatigue, and sleep problems were most strongly associated with the presence of other problems. High distress was present in 81.4% of patients reporting all 3 of these problems (DT M = 6.4). When analyzing only the subset of physical problems, fatigue, problems, getting around, and indigestion showed the strongest association with the remaining problems and 76.3% of patients with all 3 problems were highly distressed (DT M = 6.1).

Conclusions: Our results show a high prevalence of psychological distress in cancer patients, as well as a set of problems that indicate the likely presence of other problems and high distress and can help clinicians identify distressed patients even if no routine distress screening is available.

The measure of distress was crude and nonspecific, but using it is consistent with political push of International Psycho-Onocology Society to mandate screening  for distress in all cancer care settings with a distress thermometer.

The screening contains a single‐item visual analogue scale ranging from 0 (“no distress”) to 10 (“extreme distress”) to quantify the global level of distress experienced in the past week including the current day and a standardized problem checklist containing 34 potential causes of distress (yes/no questions) that are grouped into 5 categories including physical problems (20 items), practical (5), family (2), emotional problems (5), and spiritual/religious concerns (2). The questionnaire gives the following instructions: “First: Please circle the number (0‐10) that best describes how much distress you have been experiencing during the past week, including today. Second: Please indicate if any of the following has been a cause of distress in the past week, including today.” A score of ≥5 at the visual analogue scale is recommended as a cutoff for a clinically significant level of distress.


The…standardized problem checklist containing 34 potential causes of distress (yes/no questions) that are grouped into 5 categories including physical problems (20 items), practical (5), family (2), emotional problems (5), and spiritual/religious concerns (2). The questionnaire gives the following instructions: “First: Please circle the number (0‐10) that best describes how much distress you have been experiencing during the past week, including today. Second: Please indicate if any of the following has been a cause of distress in the past week, including today.” A score of ≥5 at the visual analogue scale is recommended as a cutoff for a clinically significant level of distress.

top 10 concerns distress

The problem checklist is not a research-friendly  instrument, it is a psychometric disaster. I could go on for the rest of the post explaining why, but here are some problems. It does not inquire about whether problems are present, only if they are the source of distress. There is no good way to obtain a meaningful, generalizable score because that would involve some assumptions of the relative equivalency. Early versions of this kind of measure had many more items, but produced anomalies. Like transportation and parking being the most frequent problems and, importantly financial problems. These were simply eliminated, likely because they don’t require mental health intervention, But these are the things that matter. In the Netherlands, we modified the recommendations of IPOS and asked if patients wanted help from the cancer center in addressing these problems. Most patients did not, either because they were getting help elsewhere, thought the cancer center was not the appropriate place to get help, or, mainly, thought they could handle the problems themselves.

Low and biased recruitment. The bias in which patients were recruited vs declined exceeded associations of distress with other key variables.

Out of 5889 eligible cancer patients, 4020 participated in the study, leading to a final sample of 3724 patients with complete data on psychological distress. Most frequent reasons for nonparticipation were “not interested” (n = 993, 55% of nonparticipants) and “too burdensome” (n = 588, 33%). As reported previously,32 nonresponder analyses revealed that study participants were younger (P < .001), more educated (P < .001), and more likely to be recruited from a cancer rehabilitation center (P < .001) than nonparticipants.

Not surprisingly, many patients did not want to be bothered with completing questionnaires with no clear benefit. Chrissake, these people are there to be treated for a life-altering and often life-threatening condition, cancer. Researchers should leave them alone  unless what they require of the patients clearly contributes to the patients’ well-being or could conceivably contribute to the welfare of other patients.

We found that 52% of the total sample reported clinically significant levels of psychosocial distress (≥5 on the visual analogue scale). Our one‐way ANOVA revealed that the levels of distress varied significantly for sociodemographic and medical groups. The highest levels of distress were found in women, patients who were 60 or older, unemployed, had cancers of the female genital organs or pancreatic cancer, or were in advanced stages of the disease (Table 1). Time since current diagnosis was minimally yet significantly correlated with distress (r = 0.06, P < .01).

Distinctly mental issues are not the most commonly endorsed.

The most prevalent problems were fatigue (56%), sleep problems (51%), and problems getting around (47%; Table 2). On average, patients had 8 (SD = 5.6) problems (range 0‐29). Women endorsed 26 of the 36 physical and psychosocial problems significantly more frequently than men did. Men, on the other hand, more frequently reported sexual problems as well as changes in urination (P < .05).

We found no significant difference in the number of problems among inpatient care, outpatient care, and rehabilitative settings (P = .88). The total number of problems was significantly associated with higher distress (r = 0.56; P < .001).

Cluster analysis of the problem checklist is somewhere between meaningless and misleading:

Table 3 shows the results of a cluster analysis using all 34 DT items covering practical, emotional, physical, family, and spiritual problems. We identified 2 physical problems (sleep problems and fatigue) and 1 emotional problem (sadness) as most strongly associated with psychological distress on the visual analogue scale.

Inconsistency with past studies is acknowledged but the authors ignore the bulk of the literature.

Our results suggest a lower prevalence of distress than Meggiolaro and colleagues (60%)16 but a higher prevalence than Kendall and colleagues (33%),38 using the same instrument (DT) in similar populations. Discrepancies in distress could be caused by different sample compositions in sex, age, included cancer types, and treatment stages. In particular, our study used a random epidemiological sample, while some other studies may have used a self‐selection of patients seeking psychosocial support. We also found higher levels of distress compared with an early study by Zabora and colleagues,1 who found high distress in 35% of the patients, using the Brief Symptom Inventory. However, prevalence estimates derived from different instruments are difficult to compare .

I’ll buy that.

Our sample was slightly biased toward younger age, higher school education, and rehabilitation setting.32 In addition, the MONA we used to identify problems that best represent the remaining problems is a relatively fresh approach to the analysis of symptom clusters related to distress in cancer patients. Because our analysis does not guarantee that the problems identified represent the most distressing symptoms in general but is limited to the pool of problems from which they were chosen, we also aim to pursue further studies on this highly relevant subject to compare and replicate our results by using broader sets of potential indicators. In addition, the cross‐sectional design did not permit inferences on causality. The age‐limited inclusion criteria do not allow generalization to very young or old cancer patients.

Oh gawd, more research to follow, of limited scientific value, but politically useful.

surely you jestClinical implications and conclusions (?)

Our results provide crucial insights for health‐care professionals regarding the large number of patients facing a high psychosocial and physical symptom burden. In the spirit of personalized medicine,

indicators of distress and therefore need of psychosocial support should be taken into account during routine inpatient and outpatient cancer care. We therefore identified the core problems that can indicate high distress and are easy and quick to assess. We hope that this represents a significant step toward better detection and treatment of psychosocial comorbidity in cancer patients.

This article is indicative of what you find in Psycho-Oncology, except that it is a large and adequately powered study. Much of what is published in that journal are either underpowered intervention studies published with a strong confirmatory bias or studies correlating self-report questionnaires administered to convenience samples of cancer patients with low recruitment rates. The journal is not a place to which I would turn for cutting edge, top quality research.

I have never been a member of the Psycho-Oncology editorial board, but for a while reviewed numerous articles and attempted to contribute to raising the standards for what is published there. I found that I was typically given manuscripts from big name authors that the  editors did not want to publish but sought an authoritative negative review to justify their decision. Some really bad papers came my way and I explained why they were bad. But some papers I liked and said so. I never got any editorial decisions or feedback, and some manuscripts that I thought were good simply disappeared.

I was incensed that one paper that I did not review was published with an acknowledgment for my help. The authors had rejected my suggestions and did not ask my permission to name me.  Ioana Cristea and I wrote a critique that was accepted, but then the authors objected and it was not published. You can find an account of this at Shhh! Keeping quiet about the sad state of couples interventions for cancer patients research .

Key papers are published in Psycho-Oncology without peer review, but with no notification of readers.

For a while I went to IPOS meetings. I seemed to be one of the very few outspoken critics of the organizations’ dogma, at least one of the few senior people willing to disagree with it sober and in public. Then the conference came to Rotterdam. Local organizing committees for IPOS typically get to chose a few speakers, but most slots are left to the international committee and their slots go to themselves or their cronies. The Dutch committee wanted me to debate Alex Mitchell on the merits of screening for distress. Members of the international board objected. When they failed to prevail, they tried to limit my transportation reimbursement to a round trip train ticket. I live in Philadelphia.

The debate nonetheless occurred and was lively and well-attended. The President of IPOS, who was not a speaker, interrupted so many times that I had to ask for equal time.

Acta Oncologica offered both Alex and myself $500 each for invited papers based on the debate. Alex’s paper was published with little revision, even though it greatly exceeded the page and reference limitations. Mine was given a “revise and resubmit” by a reviewer who asserted that he was biased because he had just accepted Alex’s paper with minimal revision. For me, however, he insisted on a change in tone and a downplaying of the evidence I was presenting because it conflicted with practice guidelines set by IPOS. I announced my paper was being held hostage,  and so it remains captive, well fed, but getting old.

A version of my slides was placed on Slideshare.

I got numerous invitations to re-represent the talk from senior people who were present and were offended at my treatment. I gave the talk lots of places, including St. Andrews, Hong Kong, and Melbourne. One version of my slides is here. Together with the version from IPOS, there have bee 6,000 views.

The IPOS Early Career Investigator group invited me to give a talk to them, expecting it would be closed. The IPOS President showed up and began interrupting. One of the ECRs nervously told him to be quiet, because it was my talk and he was not invited. I turned the talk into one of my most popular blog posts at PLOS Mind the Brain Junior researchers face a choice: a high or low road to success?

Slides are available here and have received thousands of hits.


I don’t go to IPOS anymore and don’t get asked to review for Psycho-Oncology. The same editors continue their extraordinary long terms as editors. I hear from insiders that their position pays wella s a result of an unusual agreement with Wiley. I heard that two years ago in DC, at the opening ceremony for IPOS, participants had to sing over and over the same few lines shown in the slide. Try to imagine the tune or try to dance to it. I heard complaints that they did not get a second free drink and that they had a miserable time, but the band played on.

psycho-oncology song