Workplace Stress Causes and Consequences

tores, rahe

A Note from Paul J. Rosch, MD, FACP:

Numerous surveys confirm that job stress is far and away the leading source of stress for adults and that it has increased significantly in recent years. But how significant or meaningful are such surveys? How accurately do they measure job stress or predict its adverse health consequences?

The validity and value of any survey depends on how, when and where it is conducted, who is sponsoring it, what questions are asked, how they are phrased, the demographic group being targeted, and a host of other factors that can be manipulated to achieve specific results. For example, an afternoon survey to home telephones sponsored by a female deodorant manufacturer concluded that housewives were under much more stress than CEOs. Surveys initiated by unions or associations can also be skewed to obtain better benefits and/or higher wages for their members.

The proof of the pudding is in the eating. What scientists want is proof that high levels of job stress increase the likelihood of future stress related illnesses. This requires defining job stress as well as a method to rate its severity that can be standardized so that data from different studies can be compared. As will be demonstrated, no one has done more to achieve this goal than Tores Theorell, Professor Emeritus of Psychosocial Medicine at the Karolinska Institute and Scientific Consultant to Stockholm University’s Stress Research Institute.

His father, Hugo, had been awarded the 1955 Nobel Prize in Medicine and his two older brothers were also physicians, so his choice of a career in medicine is not surprising. However his interest in stress research was different, and somewhat unique at the time. It was stimulated due to the influence of Dr. Richard Rahe, one of our Hans Selye Award recipients, best known for his role in developing and periodically updating the Holmes and Rahe Social Readjustment Rating Scale. This has been the gold standard for ranking the stress of various life change events and demonstrating their ability to predict future illness for almost five decades.

I had long intended to devote a Newsletter to Tores, who has participated in our Montreux Congresses, but realized that Dr. Rahe could conduct a much more meaningful and comprehensive interview, since he and Tores have been close professional and personal friends for well over forty years. Richard kindly accepted my offer and the results of their combined efforts certainly justified this decision. This is the complete manuscript of that interview, complete with charts, references and other supporting documents.  A condensed version suitable for printing, was published as the May 2013 Issue of Health and Stress, which is housed permanently in the AIS Archives.

[divider /]

RR-Hello Töres! Thanks for accepting this invitation to be interviewed as a standout scientific investigator in the field of Stress and Disease. As way of introduction, let me present brief historical information about Dr. Rosch and his monthly newsletters, emanating from his American Institute of Stress, which have previously carried fascinating and interpretive presentations of a variety of issues regarding life stresses and its effects on both animal and human physiology – often leading to illness. Among these newsletters, he has most recently conducted interviews with noted pioneers in these fields of study. Close to your home in Stockholm, Paul recently published such an interview with Lennart Levi, M.D. – with whom you had been associated for a number of years. Because of our long associations, both as co-investigators as well as close friends, Paul asked me to conduct this interview with you. Let me begin with my recollections as to how I came to meet you. I had worked for several years, first as a medical student and later as a resident physician with Thomas Holmes, M.D., reviewing his prior life stress and illness studies and later proposing new paths to explore. Dr. Holmes told me that by following this course of action I would become a “third generation” Cornell University School of Medicine researcher following the school’s tradition for such studies.

Harold G. Wolff, M.D., Professor and Chairman of the Department of Neurology at Cornell’s Medical School, was the “first generation” leader with his pioneering studies of life stress and the onset of headache. Stewart Wolf, M.D., William Grace, M.D. David Graham, M.D., Lawrence Hinkle, M.D., and Dr. Holmes became “second generation” Cornell researchers. When I readily agreed to become a “third generation” Cornell researcher Dr. Holmes insisted that I remember one important duty! That was that I should later recruit and train a “fourth generation” Cornell researcher.

As you remember, I later conducted several studies measuring subjects’ recent life changes magnitudes, constructing a series of Life Change Units (LCU) and using them to predict subjects’ vulnerabilities for developing illnesses or accidents over the following year. I greatly expanded these studies at the U.S. Navy’s Medical Neuropsychiatric Research Unit – where I completed obligated 2 years of active duty required in the United States of recently graduated physicians during the Vietnam War.

I later met many of the Cornell “second generation” researchers at the annual meetings of the American Psychosomatic Society. Dr. Stewart Wolf was particularly interested in my Navy studies of recent life change measurements for Navy officers and enlisted men developing a myocardial infarction. He later was influential in my receiving a National Institute of Medicine Special Fellowship to continue stress and cardiac disease at the Karolinska Institute’s Seraphimer Hospital in Stockholm, Sweden.

Dr. Wolf was well acquainted with Professor Gunnar Biörck, chief medical officer at this hospital. Shortly after my arrival in Stockholm, you entered the office where I was located and introduced yourself. I soon learned that you were raised in a medical family – with your father being a past Nobel Prize Winner from his research in Biochemistry and your two older brothers that had chosen a medical career. Tell me more.”

 

TT- Of course my father (and my two older brothers who became medical doctors) had a prominent role in my early life. My father certainly influenced my choice of career, although he never did so by telling me that I should become a doctor. When I had finished high school with fine grades (which made it possible for me start medical studies) I felt that studies in medicine might, in a way, “postpone” later real choices in life since the range of working tasks for a medical doctor was enormous – ranging from psychiatry to surgery, basic research, forensic medicine and general practice.

When I had finished these studies I was employed by Gunnar Biörck and he initially had the idea that I should “continue” on my father´s path, which meant doing biochemical research within internal medicine (Prof. Biörck had done his dissertation in biochemistry with my father as his supervisor). I did not feel comfortable with that and I realized that this construction was unrealistic – I would have had to focus 150% on biochemistry and I wanted to focus instead on internal medicine (which by the way also meant 150%). After some months I told Biörck that biochemistry was not my cup of tea, but I also told him that I was very interested in psychosomatic medicine after I had read Stewart Wolf´s edition of Harold Wolff´s famous book Stress and Disease. Gunnar Biörck was also quite interested in psychosomatic medicine himself and it so happened that Dick Rahe was scheduled to spend two years with us. So Biörck simply said that I should work for him (you).

This is how stress and psychosomatic medicine started for me. Personally I believe that this was exactly the right moment for this kind of junction. There was a beginning interest in “ soft questions” in internal medicine but we had to translate those into numbers and statistics in order to be accepted by our colleagues (Life Change Units were perfect). And the need to rely on numbers is of course still true to a great extent today, although there is a wider acceptance now of soft values.

To return to my father, he was a wise man so he never criticized my choice of scientific topic but of course he was a true biochemist so when I was a child I remember him saying that those who have to prove anything by means of statistics are “not really serious scientists”. Of course, I have been rather heavy on statistics in most of my scientific endeavors.

 

RR-Tores, I clearly remember you were somewhat shy about asking to work with me! None-the-less, you soon began suggesting some studies of Swedish cardiac patients who were receiving ongoing follow-up care at the hospital. We first agreed to pursue the question: “Do patients recovering from a heart attack remember more life change events over the year or two prior to their attack, compared to life events that occurred a few years earlier in time, because recent memory is better than remote recall?”

To explore this question you gathered a group of 54 post-infarction patients that had been followed for several years following their infarctions and asked them to list their life changing events over two to three years before their heart attack (scored as LCU) along with recent life change events (also scored for LCU) covering the recent 2 to 3 years following their recovery.

The results for the 27 patients who had had their infarction two years before the interview are shown in Figure 1 below. They illustrate a significant LCU build-up over two years preceding these patients’ infarctions, with significantly lower LCU levels across two years of their recent lives following recovery. This study remains a unique answer to prior questions raised concerning the validity of reported pre-infarction LCU build-up being possibly due to a superiority of recent recall compared to more remote recollections. 27 patients who were hospitalized for their infarction and had had no previous infarction reported a continuous buildup of life changes during the year preceding their myocardial infarction (not shown).

Rahe Tores Figure 1

 

TT- In addition, we did something that was quite original and has

not been published, in detail, in a scientific journal. We compared three pre-infarction years of life change scores from 20 patients to the same three years of life change scores provided by their spouses. Spouses had been instructed to use their own recollections and not ask for their husbands’ input when completing their own lists.

Once gathered, we calculated correlation coefficients between husband and wife recent life changes LCU scores. We found the highest coefficient, 0.69, for the year immediately prior to the husband’s heart attack. Coefficients for the second and third years prior to infarction were 0.57 and 0.56. We were impressed by these unexpected high values, as husbands may well not tell their wives about some recent life change events they were not happy about experiencing.

 

RR- Tores, I did reference the study you just mentioned supporting the validity of spouse reports concerning a husband’s recent life changes in a publication of an investigation I carried out with Evy Lind, MSW. She identified 39 males listed in Stockholm’s most recent yearly registry for males dying outside of a hospital later determined as caused by a myocardial infarction. Spouses of these men, and one child, (83% of the requested individuals) agreed to provide a listing of the victim’s pre-death life changes from our survey instrument for the three years prior to the victim’s death.)

Of this sample, 14 victims had a previous history for coronary disease while 25 men had no such medical histories. For the 14 victims with prior coronary disease, life changes total for the year prior to death (210 LCU) was twice the magnitude as LCU totals for the second (110 LCU) and third (95) LCU) pre-death years (p<0.01). See Figure 2 below.

rahe tores Figure 2

For the 25 victims without prior heart disease histories, life changes magnitude for the year prior to death (230 LCU, of which 160 LCU occurred over the final six months) was three to four times higher than magnitudes for the second (70 LCU) and third year’s (60 LCU) totals (p<0.001). See Figure 3 below.

 

rahe tores figure 3

 

TT-The most original contribution to my medical thesis was my longitudinal study of 21 male patients that had suffered a myocardial infarction two years before my study. Two of the subjects had retired after the myocardial infarction but the other ones were working, the vast majority full-time. These subjects were followed week after week and a standardized life event interview was performed on a constant weekday.

Urine was collected for the assessment of catecholamine output during the workday (from waking up to coming home from work). The study enabled a calculation of the individual correlation between life change units (LCU) during a seven-day period and the urinary adrenalin and noradrenalin output during the last day of that period.

The findings on adrenalin were interesting. The results showed that one-third of the subjects had strong relationship between life change units per week and adrenalin output and another third showed a moderate relationship. The remaining one third of subjects indicated no relationship between life change units per week and adrenalin output.

Table 1 below shows the linear regressions for each one of the subjects and figure 4 all the values referred to the individual´s own means for life changes units/week (x axis) and adrenalin output (y axis). A crude approximation is that 100% increase in life change units during a week corresponded on average to a 50% increase in urinary excretion of adrenalin – with large variations between individuals.

As far as I know there have been no replications of this study. However, a striking observation was that the real life variations in adrenalin output that we observed were often more pronounced than the changes we have later observed in our laboratory experiments with humans.

 

Tores, rahe 4

Linear regression coefficients for each patient in the week-to-week life change/adrenalin study are shown in Table 1. The coefficients show increases in urinary daytime adrenalin output (nanograms/minute) corresponding to one LCU increase in life change sum in a defined week.

 

Rank order                          N studied week                  Linear regression

1                                              3                                              2.4

2                                              7                                              0.7

3                                              7                                              0.6

4                                              9                                              0.5

5                                              6                                              0.5

6                                              7                                              0.4

7                                              6                                              0.4

8                                              9                                              0.3

9                                              4                                              0.3

10                                           2                                              0.2

11                                           7                                              0.2

12                                           8                                              0.2

13                                           14                                           0.1

14                                           7                                              0.1

15                                           4                                              0.1

16                                           8                                              0.1

17                                           10                                           0.0

18                                           9                                              0.0

19                                           9                                              -0.1

20                                           8                                              -0.1

21                                           6                                              -0.2

Mean regression                            0.320 (p<0.02)

 

Table 1.

Figure 5 below shows an individual case, a 49-year old foreman at a mechanical firm. The work events that he reported during weeks 1, 4 and 9 all had to do with conflicts either with his boss or with work mates. The family event reported in week 2 was related to his work situation and mirrored a tension between his work and family situation. During week 7 he reported that a close friend had suffered a myocardial infarction. A series of “family” and “personal” events taking place during weeks 8 to 11 were related to his mother falling ill and going through surgery as well as conflicts arising due to this. High levels of epinephrine excretion are observed particularly during his mother´s illness episode. Episodes of angina pectoris were reported during weeks 7 and 8 (week 8 was also the week with his highest adrenalin output).

 

Tores rahe 5

 

RR– Very nice work Tores! I’ve used your Figure 4 occasionally when patients in my practice report experiencing angina pectoris along with reports of major life stresses in their lives. I point out the last line in the Figure for weeks 7 and 8, showing that angina pectoris occurred for this individual as he experienced a cluster of newly stressful life changes. I then add that serum cholesterol even rose during these weeks. I had suggested that you to add this biochemical measure to your study as I had previously found marked elevations in serum cholesterol for persons who were feeling overwhelmed by their current life stresses.

Below is my most illustrative case history shown in Figure 6. I was practicing Liaison Psychiatry at the time that involved being called to see hospitalized medical patients showing some worrisome psychological problems. In this case I was called to evaluate a newly admitted patient to the Coronary Care Unit (CCU) following a heart attack. This individual couldn’t calm fears of dying over his first days on the Unit, spending hours carefully listening to electronic beeps coming from his heart monitor with each of his heartbeats. These beeps reassured him that his heart was still functioning. His fears slowly subsided over several days on the Unit. But they returned when nursing staff told him he was about to be transferred to the General Medical Ward. Leaving the CCU he would no longer be able to listen for a cardiac monitor’s reassuring beeps.

In trying to build confidence for the transfer, I pointed out to him that his daily blood draws showed progressive improvements in serum cholesterol, indicating he was slowly recovering! This information seemed to help him accept a transfer to the General Ward. On the day of transfer there was a large spike in his serum cholesterol! But the following week his cholesterol levels once again begin to fall. See figure 5.

Then his wife came to visit and as it also was his birthday she brought his favorite cake – one made in large part from whipped cream! However, doctors and nurses had already told him that eating high cholesterol foods could cause another heart attack! He needed to be readmitted to the Coronary Care Unit for four days after having just one slice of cake.

On my continuing visits with him I pointed out that his cholesterol went from 225 mg % to 174 mg % over his final two weeks on the General Ward. His fears of death had ceased and he looked forward to going home.

tores rahe 6

 

Tores, I recall that after you completed your thesis work you achieved positive working relationships with two very highly regarded stress researchers – Stewart Wolf, M.D. and Lennart Levi, M.D., PhD. Tell me how these associations came about and how they helped shape your budding research career.

 

TT- The opponent of my doctoral thesis was Stewart Wolf who was a good friend of Gunnar Biörck. It was of course a fantastic experience to have such an eminent opponent. Stewart Wolf became genuinely interested in my work and invited me to become a post-doctoral fellow with him in the Medical Branch of the University of Texas in Galveston. I accepted, and when I was in the United States working with Wolf, Lennart Levi contacted me and offered me a position at the Stress Research Laboratory (which was part of the Karolinska Institute). I had already collaborated with Lennart using his laboratory for my catecholamine analyses used in my thesis work. Therefore, to continue my research work with him seemed a natural progression.

However, the months (almost a year) I first spent with the legendary Stewart Wolf in Galveston were very important in later shaping my career. He offered me the possibility to analyze data from a large sample of myocardial infarction patients he had gathered while serving as a Professor of Internal Medicine at the Medical School in Oklahoma City, Oklahoma. These patients had been followed individually for up to ten years. Among other cardiac tests that he had carried out were ballistocardiographic recordings, performed every other month, coupled with interview data covering these patients’ ongoing life events.

Ballistocardiography is a method for recording contractile patterns of the heart. The patient is placed on a recording table and sagittal bodily accelerations are recorded. As you recall, I collaborated once again with you and we published our findings in the Journal of Human Stress (Theorell T and Rahe CR 1:18-24, 1975).

A summary follows.

tores rahe 7

 

Of 36 men and women that had experienced a documented myocardial infarction, half of them ultimately died from their disease over a 6 years of follow-up while the other half survived, providing longitudinal recent life changes and ballistocardiographic data. The 18 patients who had died indicated in their interview data a significant buildup in life changes that peaked approximately one year prior to death. During this time frame their serial ballistocardiograms showed a significant buildup in average forces of contraction which peaked approximately six months prior to death. The 18 post-infarction patients that had survived six years of follow-up showed neither a buildup in life changes nor an increase in the ballistocardiographic index of cardiac contraction force. These findings of a life changes peak preceding ballistocardiographic evidence of an “overworked” heart were discussed in terms of their possible medical significance.

 

 

 

Accordingly, these observations indicated that there was a typical time sequence preceding coronary death in these patients: First a build-up of critical life change events followed by an increase in cardiac contraction force, leading to increased metabolic demands and oxygen needs in these patients with narrowed coronary vessels. Another article from the same study singled out those patients who lacked life changes events data. In this study there was a control group without coronary heart disease. Here it was shown that patients who died during the study showed poor contractility, and during the 6 months preceding death the contractility increased significantly. (TT, Dan Blunk and Stewart Wolf, J Lab Clin Med 86:46-56, 1975).

Stewart Wolf was one of the most important researchers in psychosomatic medicine at the time and he organized a series of visits for me with a number of prominent psychosomatic researchers including George Engel, Lawrence Hinkle, Arthur Schmale, Ray Rosenman and George Kaplan. During that year I also visited you and your family in San Diego.

 

RR- I remember your visit well. We worked on preparing the publication you mentioned above. Now tell me how your career progressed working with Lennart Levi.

 

TT– When I came back to Sweden and started working in Levi´s laboratory I combined this effort with clinical work in Biörck´s department of medicine. I did not want to leave clinical medicine entirely, so I still participated as a “real doctor” and even took call at the hospital. The department had purchased a ballistocardiograph allowing me to continue studies of variations in contractility for patients with heart disease.

Several studies were carried out using ballistocardiographic assessments during stress interviews, following Stewart Wolf´s tradition. In addition to the contractile patterns variations in blood pressure, peripheral circulation, ECG and plasma growth hormone were also studied.

The adult twin brothers that were discordant with regard to coronary heart disease, that you had previously studied, were studied during stress interviews using all these assessments. One pair member always showed far more advanced coronary heart disease than the other twin brother. Even among the other monozygotic twin pairs in the study, we found a few interesting dissimilarities between the members of a twin pair, although similarities in cardiovascular reaction patterns among these twin pairs were also very prevalent.

In Levi´s laboratory I conducted epidemiological studies using life changes measurements as my main perspective. One investigation was a prospective study of life changes in relation to myocardial infarction risk. Subjects were 8000 building construction workers who were examined in public hospitalization and death registers, two years following the life change questionnaire being distributed and completed. We asked whether specified life changes had occurred during the past year and if such events were related to subsequent illness risk. Although it was not possible to examine for a recent build-up of life changes as we had in earlier studies, we could determine topics of particular relevance for specific illnesses. For instance, critical changes at work (such as increased responsibility and conflicts) were found to be associated with increased risk for developing a myocardial infarction over the following two years. (Theorell, T., Flodérus-Myrhed, B.: “Workload”and risk of myocardial infarction – a prospective psychosocial analysis. Int J Epidem 6(1):17-21, 1977.)

 

RR– You now had academic degrees and advanced training in research, but you continued to extend your academic training into new disciplines. Tell me about these transitions.

 

TT– As I was involved with these field experiments examining subjects’ metabolic and hormonal changes in light of various life adversities such as voluntary starving and changes in shift work schedules, I felt I needed further training in epidemiology. Therefore, I chose studies in the Department of Social Medicine at the Huddinge hospital (that is part of the Karolinska Institute academically). There, the dynamic professor Erik Allander had recruited statisticians well trained in mathematics. In addition, there was an emphasis on philosophy and the development of illness concepts.

During the two years working and teaching there, I met Robert Karasek who had just introduced his demand-control model for studying job stress conditions. He was aware of my interests in the relationship between job conditions and risk of myocardial infarction. We soon began our extensive collaboration, and I suggested using Swedish cohorts to study the relationship between high psychosocial demands in the presence of low decision latitude at work (a combination which was labeled “job strain”), on one hand, and risk of myocardial infarction on the other hand.

 

RR– The measures you and Karasek developed for these studies became quite well known and resulted in your eventual collaboration with many other medical and psychological researchers. You also became a candidate for Lennart Levi’s position as the Head of the Karolinska’s Social Medicine Department. I was invited by you to visit Stockholm to serve as a Medical Thesis Opponent at this time. You had been remarried a few years before, moved to a new home in Stockholm and were raising two fine young sons. These recent life changes coupled with the challenges of competing for the top position at your work made you a candidate for life stress and illness. Can you say a bit about how you fared during this trying time?

 

TT– When Karasek and I started collaborating, our joint questions had to do with the possible epidemiological association between job strain and myocardial infarction and we also wrote a theoretical paper on possible physiological mechanisms behind such an association. After more than ten years of intensive collaboration, our book Healthy Work came out in 1990 (Basic Books, New York). This has been widely quoted and used in universities in many countries.

Over the years these themes were picked up by many groups of researchers over the world. Several studies (starting with healthy subjects that were questioned about their job conditions) were followed prospectively with regard to their development of a myocardial infarction. Most of these studies showed job strain increasing prior to an infarction along with increasing cardiac risk factors. Ultimately, several European groups collaborated (Kivimäki et al Lancet. 2012 Oct 27;380(9852):1491-7) in an extremely large prospective study comprising 197,000 subjects. Their results showed job strain increases that were independent of other cardiac risk factors. The association was found for both for men and women.

Physiological mechanisms have been illuminated in many studies (blood pressure regulation, catabolic and anabolic hormones, immune reactions). During later years my group has also been studying possible job interventions for improving the conditions and we have also collaborated with researchers testing other job stress models (such as the effort-reward model that had been introduced by Johannes Siegrist). Finally there is also a growing research on adverse mental health consequences of poor psychosocial job conditions in the world, and a substantial part of this research is based at least partly on the demand control model which in its extended form includes social support at work, a dimension that was operationalized and put in theoretical context by Jeffery V Johnson, an American sociologist who spent working time with us (together with his wife at the time, Ellen Hall, also a successful sociologist) in Stockholm.

The life event perspective that you had introduced in my scientific life has remained as an important topic in my scientific life. It was fortunate for me that you came into my scientific life. The clinical utility of the life change concept is striking. In addition one dimension in critical life changes is “lack of control” (close to low decision latitude) that has been an overriding concept in a lot of my research throughout my whole career. The social medicine perspective and the perspective that Karasek brought into my picture have made me aware of the importance of societal perspectives.

After the two years at the department of social medicine including three months

at the Columbia University in New York working with Karasek I was appointed full professor of health care in 1981 at Lennart Levi´s newly started National Institute for Psychosocial Factors and Health, a researching governmental agency under the ministry of health. This institute has been closely connected to the Karolinska Institute. During the first years I was also serving as a chief physician for psychosocial medicine at the Karolinska Hospital as part of my job at the National Institute. When Levi retired in 1995 I was appointed his successor and at the same time I became a full professor at the Karolinska Institute.

The collaboration between Karasek and myself was occupying considerable effort and time and this was not good for my marriage. And (although I did not know this) I had a competitor, an author who is a well-known member of the Swedish Academy (for literature). After twenty years of marriage my first wife and I separated. My children Tobias (who is now a stage director of theatre and opera) and Ebba (who is an education expert with training in the production of child movies) were in the upper teens and of course this was a difficult period for all of us. I married my present wife in 1987 and in the second marriage I have two sons. The older one, Jakob, is a medical student at the Karolinska Institute and he is also a doctoral student in immunology. The younger Axel is studying technical physics at the Royal College of Technology. Since work continued to be very intensive these years were difficult. I worked very long hours. On a typical working day I started working in my office at 7 am, arriving home mostly at 6:30 pm and then continuing to work for an hour after the family had gone to bed. Fortunately however, all of us in my family (both the old and the new family) were active in music (I sing and play the violin myself) so we had great joy in making music in various combinations over the years. This helped (and helps) us all.

During later years I have been doing research on health effects of cultural activities. For several generations in my maternal and paternal background music has been a very important component and I have been driven by efforts to explain why some music activities may be used in health promotion and to answer the question whether there are “real” effects and not only imagined ones. For instance in a random control trial study we showed that irritable bowel syndrome patients who started singing in a choir for a year had a better development of plasma fibrinogen and also tended to have a better pain development and development of the hormone motilin (important for bowel function) than a control group who had group lectures and discussions. In addition, during the first half year the choir group had a higher secretion of testosterone in saliva, an index of improved regenerative activity in the body (Grape, C., Theorell, T., Wikström, BM., and Ekman, R. Choir singing and fibrinogen. VEGF, cholesystokinin and motilin in IBS patients. Medical Hypotheses. 72; 223-234, 2008. and Grape, C., Wikström, B-M., Ekman, R., Hasson, D. And Theorell, T. Comparison between choir singing and group discussion in irritable bowel syndrome patients over one year: saliva testosterone increases in new choir singers. Psychother. Psychosom. 79: 196-198, 2010). In other studies we have shown that the flow experience is characterized physiologically by high sympathetic activation (arousal) and at the same time a high parasympathetic activation as it is mirrored in heart rate variability (Manzano, Ö., Harmat, L., Theorell, T. And Ullén, F. The pathophysiology of flow during piano playing. Emotion 10(3), 301-311, 2010 and Harmat, L. and Theorell, T. Heart rate variability during singing and flute playing. Music and Medicine. 2(1), 2010.)

A recently published study was on the effect of manager education based upon art experiences. The goal was to make managers more engaged in their employees´ health and psychosocial conditions. Other research has shown that the worst manager behavior from the employee health perspective is the passive withdrawn one. Two stress improvement training programs (both lasting for nine months) were compared in a random control trial. One program (Shibbolet) consisted of monthly sessions with poetry reading activating ethical choices interspersed by music selected for the situations. Group discussions and diaries followed each session. The comparison program was a more conventionally designed high quality program (with lectures and discussions) for improved psychosocial management. For comparability there was equal time commitment in the two groups. The managers as well as their subordinates were followed for 18 months after start and the results showed that both the managers and the employees benefited more from the artistic manager program than from the more conventional program. Psychological health and plasma concentration of dehydroepiandrosterone sulphate (DHEA-s, a hormone with regenerative/reparative and stress protective effects) developed more favorably in the shibboleth/manager group´s subordinates. (Romanowska, J., Larsson, G., Eriksson, M., Wikström, B-M., Westerlund, H. And Theorell, T. Health effects on leaders and co-workers of an art-based leadership development program. Psychother. Psychosom. 80; 78-87, 2011.) In a recent publication we have shown that work sites with more cultural activities organized for the employees are protected against unfavorable development of burnout (or rather emotional exhaustion) than other work sites. Accordingly our results seem to confirm that cultural activities at work may have a protective role for the employees (Theorell T, Osika W, Leineweber C, Magnusson Hanson LL, Bojner Horwitz E, Westerlund H.: Is cultural activity at work related to mental health in employees?

Int Arch Occup Environ Health. 2013 Apr;86(3):281-8.)

I am presently involved in a large project on the development of music interest and pursuits for twins, again using the Swedish twin registry. I have some responsibility for the health parts of the investigation of 8000 individual participants. The twins are asked in detail about music in their environment, about music teaching and hours of training during different parts of their lives. They are exposed to objective musicality tests and there are detailed questions about health development and alexithymia (TAS 20). Results from this study might enable us to draw conclusions about the role of the environment (and of the influences of genes between identical and fraternal twin pairs) in the development of musical and other culture related skills, emotional competence (the opposite of alexithymia) and their effects on health promotion and resistance to illness.

You and I have kept in contact over the years and I will always be grateful that you jumpstarted my scientific career. I became godfather to your daughter Annika who was born in Sweden, and. despite the fact that she has lived most of her life in California, presently as an English teacher, she has learned Swedish. She and I have exchanged letters in Swedish and last year she visited us in Stockholm with her two children. I look forward to many more years of close association and greatly appreciate this opportunity to express my indebtedness to you.