I am interested in how often clinicians are aware of the interaction of medical and psychiatric conditions. Can anyone offer their personal experiences in this area of practice. Also, I would welcome references for publications that address the interaction of medical and psychiatric treatment.

This is a little more complicated than it looks. A pathway analysis done on the Australian MH Survey by Goldberg and others suggests that there isa meta structure for disorders of emotional distress into externalizing disorders (where you act out) and Internalizing disorders. The internalizing disorders divide into distress (depression, GAD), fear (panic and phobia) (1).

If these authors are correct -- and I think they are -- then the comorbidity we see is a consequnecce of out meta structure and the tendency in DSM to divide psychiatric syndromes finely. This sis supporedt by the very high rates of co occuring anxiety disorders and co occuring anxiety and mood disorders (which, at least in the NZ survey and tmost of the otehr World Menalt Healt Intiatve survesy, is greater than co occuring substance misuse disorders).

At present we have to accept that there will be co occuring diagnoses. The clinically sensible way around this is to formulate the problem (why this syndrome now?) and then note the principal or presenting syndrome and how the other syndromes contribute to this.

If you are interetsted in medical=psychiatric co-,morbidity, there is a world mental health intiative working group on this issue. I'm attaching some NZ and international references.

1. Goldberg DP, Krueger RF, Andrews G, Hobbs MJ. Emotional disorders: cluster 4

of the proposed meta-structure for DSM-V and ICD-11. Psychol Med. 2009

Dec;39(12):2043-59. doi: 10.1017/S0033291709990298

2. Scott KM, Kokaua J, Baxter J. Does having a chronic physical condition affect

the likelihood of treatment seeking for a mental health problem and does this

vary by ethnicity? Int J Psychiatry Med. 2011;42(4):421-36.

3. Scott KM, Oakley Browne MA, McGee MA, Wells JE; New Zealand Mental Health

Survey Research Team. Mental-physical comorbidity in Te Rau Hinengaro: the New

Zealand Mental Health Survey. Aust N Z J Psychiatry. 2006 Oct;40(10):882-8.

4. Scott KM, Von Korff M, Alonso J, Angermeyer MC, Bromet E, Fayyad J, de

Girolamo G, Demyttenaere K, Gasquet I, Gureje O, Haro JM, He Y, Kessler RC,

Levinson D, Medina Mora ME, Oakley Browne M, Ormel J, Posada-Villa J, Watanabe M, Williams D. Mental-physical co-morbidity and its relationship with disability:

results from the World Mental Health Surveys. Psychol Med. 2009 Jan;39(1):33-43.

doi: 10.1017/S0033291708003188.

You may be interested in "Global Perspectives on Mental-Physical Comorbidity..In the WHO World Mental Health Surveys" . Globally, the majority of people with chronic physcial conditions did not report anxiety or depression in the year preceding their illness, but childhood adversities and early onsel anxiety and depression were non specific risk fatcors for a number of physical illnesses. The costs of physical illness in terms of mental ehalth was much greater in the developed world. Suggests to me that a lot of the burden of mental illlness related to physical conditions in the West is associaetd with culturul/psychological factors.

I apologize for my English. Your question refers to the differential di-agnosis of behavioral disorders. It is quite complex and interesting.

First we need to determine level of the mental disorder: psychotic or non-psychotic. Behavioral disorders are non-psychotic level of reaction is most often in the practice of psychotherapy. Then we need to differential di-agnosis of a neurosis, personality disorders and acute psychogenic reaction.

Each of these states has its own criteria.

The masked (I mean the mask of somatic disorders) behavioral disorder characterized by:

- The absence of objective signs of physical illness.

- Frequency (seasonal) manifestation of painful symptoms.

-The slave being daily rhythm with the improvement in the evening, sometimes in the morning

- Persistent requests for medical assistance despite the apparent lack of treatment

- Improvement in patients receiving antidepressants

A great discussion question. Unfortunately, terribly complex. WHO-WMH surveying indicates a strong correlation between mental disorders and certain physical disorders. For example, diabetes mellitus, is known to occur more often in patients diagnosed with psychiatric diseases than in the general population. However, there is a causality dilemma as both involve immune/inflammatory dysfunction. Similarly, our work looks at mood dysregulation during and after pregnancy. Pregnancy presents a profound immunologic/inflammatory process and we've been finding strong relationships between certain biomarkers of inflammation (e.g. il-6) and maternal depression/anxiety. Personal experience suggests that practitioners outside of neurology and psychiatry simply don't want to deal with psych issues (not because they don't care, but because they don't know how to handle it) and therefore don't attend to issues of co-morbidity.

We have examined this issue among older adults with HIV, finding that the overall level of morbidity as well as specific disease conditions are related to higher levels of depressive symptoms. This is not exactly what Neil was inquiring about as our morbidity info was self-report and we examined depressive symptoms vs. a clinical diagnosis of depression. If you are interested in this paper, the citation is:

Havlik, R. J., Brennan, M., & Karpiak, S. E. (2011). Comorbidities and depression in older adults with HIV. Sexual Health, 8(4), 551-559. DOI:10.1071/SH11017

In my opinion, after 30y. of studying the subject, depression is a nero-biochemical illness with multisystemic features , especially pain, digestion, micturition, neuro-vegetative dysfunction etc.

I have performed a large study proving the association of bodily signs and traditional psychic ones. More over, both dissapeard independently one after the other, with an efficient treatment.

Therefore, it can be concluded that the so-called depression is the common cause of psychic and bodily features, which explains clearly the co-variation and offers tremendous therapeutic prospects.

Moreover, taking in account bodily signs, allows to link to depression other psychic syndroms like obessive-compulsive and certain psychotic presentations.

It to be underlined that it was partly known in the 70th but progressively forgotten since.

Finally, never forgott that depression diagnosis is blurr, taking into account the multiplicity of scales, quantitative diagnostic cut-offs and the absence of definition of non-depresion.

If you are interrested, I suggest my site http://bmaroy.free.fr chapter Depression, the second part being in english. It is the abstract of my book in french : La dépression et son traitement : aspects méconnus, L'Harmattan, Paris, 2011. In this book, it is possible to find a very large bibliography.

Some fabulous answers here. If it helps I have documented a long-term cycle in admissions for certain medical conditions among which are a group of mental health diagnoses. For more details of this research go to the 'Emergency Admissions' page at www(dot)hcaf(dot)biz This cycle appears to behave in a similar way to an infectiousn outbreak. Oh that life were simple!

The links between long term conditions (physical) anf psychological/mental health problems are now well established. It is however a complex area but there is a wealth of literature literature out there on the subject. My clinical work over the last year has focused on this group, CHD, COPD, diabetes, Renal. Significant numbers have associated depression and or anxiety.

Check out the Scottish and English MH strategies, LTC paper Kings Fund, The Living Better Project (scotland) and a really interesting paper by Barnett et al in the Lancet May 12 2012 p 37-43.

Psychotherapy in conjuction with antidepressants can have a major positive therapeutic effect, significantly improving quality of life.

Absolutely great set of responses from all of you! You've given me some great leads in my own research. I don't want to curtail any further responses, but just let everyone know how much I appreciate your perspectives and expertise.

It is important do differenciate causes of depression like pancreatic cancer or C hepatitis of its consequences like pain or functionnal dieases.

Key point about the Barnett et al Lancet paper is the association between multimorbidity (and physical and mental health co-morbidity in particular) and socioeconomic deprivation: multimorbidity happens 10-15 years earlier in most deprived decile compared to most affluent, and with considerably more mental health co-morbidity. This may be due in part to higher rates of behavioural risk factors for chronic physical conditions (smoking, poor diet, physical inactivity, alcohol excess) in more deprived areas (and among those with mental health conditions).

WHO data indicates that 30% patients who attend any general medical facility do have behavioral emotional causes. Seeking treatment is it self a behavioural trait, who and when will seek medical help depends on attitude along with other plethora of factor apart from physical illness. Stigma and nonavailability of insurance for mental disorder force many persons with mental disorders to be admitted with physical diagnosis.

To me, the answer should be articulated in three points:

1) Methodological/Research perspective: when you are exploring the relationship between two or multiple variables (e.g. one/multiple medical or behavioral conditions or their prevalence rates), you are actually doing a "correlation analysis", which be definition provides just a description of the rate of occurrence of a given or multiple events (e.g. prevalence rate of one condition in relationship to another one). So that you have a positive or negative "correlation" between two (bivariate) or multiple (multivariate) factors. But the interesting point is that correlation analysis do not give any sort of directional hypothesis (meaning that you can "read" that correlation in "round-way" just alternating "positive" or "negative" descriptions) and that you have no explanation (predictive hypothesis) or what factor should predict or drive one/more than one others. Also, a "third" variable problem should occur (meaning that you can't know "a priori" if you are controlling for additional factors which you may know or not know (yet) but that should still influence the correlation between two or multiple factors.

The only way to "control" for more factors is to perform a regression analysis, hopefully by implementing a model that is able to account for all "confounding" factors, if ever know. While in pharmacological trials a very good approach is to design a double-blind controlled (usually placebo or multiple arms) study (assuming the sample size is sufficient to allow for multiple comparisons without reducing the potency of the study), it is much more difficult to this in ecological or epidemiological naturalistic studies (prevalence studies for example).

Back in the 1950s, the preliminary research results on the correlation between the prevalence of lung cancer, caffeine consumption and smoking, lead to the apparent conclusion that coffee had a causative role in determining lung cancer. Which is not. The problem there, was that no Pearson, Spearman's or Kendall's tau regression was performed in order to reduce the chance of additional confounding factors, while, in real world, most people addicted to smoke, often is a hard-to-die coffee drinker. No more, no less.

So, about your question, who shouldn't expect to have an answer about the causative or directional effect of behavioral disturbances on medical disorders or vice-versa, simply because no correlation analysis is per se able to deal with this.

2) Biological perspective: I agree with Dr. Silverman here. The HPA axis and the neuro-endocrine-immune system is the common substrate for "medical" and "behavioral" conditions, so that even the distinction between the two makes no sense to me. Maybe, you should say that most "medical" disorders may also manifest with behavioral features (not a co-morbidity actually) and that all MEDICAL disorders should also be influenced by epigenetic factors, indipendenly on the presence of behavioral symptoms at a given point of the course of the illness.

3) A clinical perspecitve: you were mentioning mood and anxiety disorders. Indeed they are quite different if you are accounting just "simple" non-comorbid ones. By definitons, anxiety disorders follow a sort of "chronic", (almost "processual"-like) natural couse; when this is not the case, usually a medical or affective conditions is influencing the longitudinal presentation of the anxiety disorder(s). Mood disorders, are usually characterized by an episodic presentation. This has major implication for your research about the topic you are discussing here. In fact, you should ideally assess the continuous spectrum of affective disorders following a lifetime approach rather than an orthogonal "hic et nunc" one, which is the one that one should mistakenly follow focusing just on the hospitalization period of the patient (no matter the medical or "psychiatric" cause of the hospitalization). While there are relevant temperamental and trait states which are almost accounted as "candidate endophenotypes" for mood disorders (in the sense of stable personal features which are non influenced by the current episode/polarity of illness, medications and that are present and stable even in non-affected relatives), other features are not that stable. Cognitive status is also very relevant, especially in diabetic patients. And, clearly, bipolar patients with endocrine disorders (e.g. type-I diabetes), are often cognitively impaired in the long-term course, which further increase the risk of unreliable medical/behavioral recall history. Recall bias is also an intrinsic issue of BD, because most patients fail to recall previous (hypo-)manic episodes because they are "nostalgic" of missing "hights", which are the status the expect to experience again and that they consider "normal", and also because we live in a "depressocentric" world, where depression is "bad" and hyperactivity, mania and productivity is good (and also we live in "antidepressant" world by that consequence) [for reference, please see the work of Kleman or Ghaemi].

The topic is indeed complex, and to sum up, I would just recommend you to assess the prevalence and "correlation" of the factors, with no "demand" for explanation or causation. Also follow a lifetime, naturalistic approach. Then, if you are planning further explorations, try to implement a multivariate model, then a hypothesis-directed one-tailed model, and, finally, a factorial/cluster analysis.

By the way, I'm attaching here for your convenience a paper I published in the Journal of World Psychiatry about the "prevalence" of different newly-diagnosed endocrine disorders (including diabetes) in a large sample of major depressed women. As you'll notice, it is a very "plain" analysis, with just a prevalence report. Also, you'll see that most endocrine and anxiety disorders (e.g. panic for type-diabetes) largely co-occur. But please remember, when you are finding a lot of multiple co-morbidities or co-occurring phenomena, you are not actually reaching a good finding, rather you are probably witnessing at multiple phenomena sharing a common background or predicted by a common diathesis. So move on on further exploring your data... On the contrary, if you're going to search the community sample for the relationship between shoes size and depression, who may find that most depressed men may have 8.5 USA size. Maybe with a very strong positive correlation factor. But indeed this make no sense unless you're able to explain what you found and to further investigate that finding to discriminate for confounding factors!

Take care, Michele Fornaro, MD.

I've rephrased your question (correct me if wrong): Do behavioural condition hospitalizations influence "medical" condition recording rates. Leaving aside common causality and concentrating on this as a measurement effect, this requires a mechanism (say psych admission leads to testing and diagnosis). From a UK perspective this is going to be a small effect. Psych admissions are rare, diabetes is common. Even if relative risk for diabetes is incr in group also at risk for psych admissions, the absolute effect will be small (negligible). Other countries / populations may differ.

Thank you for raising the question and for all the debate.

I am not sure I entirely understand the initial question and it appears that people have interpreted the question differently based on their own research/clinical practice. As Terry mentioned in the UK, and elsewhere, the link between LTCs and mental health conditions has been well established - good examples of policy papers/discussion come from the King's Fund organisation. The acknowledgement of this multimorbidity and the impact it has on patient outcomes and services/financial provisions has lead to the UK government developing specific policy to tackle the issue 'no health, without mental health'.

In my recently completed PhD research I chose to investigate anxiety in heart failure patient populations; exploring reasons for variance in the prevalence of anxiety in this patient group and examining the amount of variance in health related quality of life scores accounted for by levels of anxiety (as measured with the HADS). Anxiety in heart failure patients is common - reported rates varied from 6%-74%. In a systematic review I found that the primary reason for the variance was the manner in which anxiety had been defined and measured. Levels of clinical anxiety were akin to levels in the general population. In a cross-sectional survey I did not find that anxiety accounted for a significant amount of unique variance in HRQoL scores; however, depression did. I feel that one reason for this may be related to the type of sample involved in my research and the manner in which they received their care – intensive 30 minute appointments every 2 weeks from specialist HF nurses; that allayed a lot of fears for patients and allowed them to talk about their condition regularly in depth.

Interestingly, from the PhD it has become apparent that the way anxiety/depression is measured will have a dramatic influence on findings. As many somatic symptoms of depression and anxiety overlap with those from LTCs, cardiovascular in particular, it is essential to measure emotional symptoms in a careful and considered manner. I am very keen to learn more about the overlap between physical conditions and emotional symptoms, as this is not an area of expertise for myself. Furthermore I am keen to learn more about how researchers are coming to understand more about anxiety and depression through research that looks at these conditions as co-morbid to other physical health complaints.

Currently I feel we do not know enough about the causal relationships between LTCs and mental health conditions (anxiety and depression), particularly in heart failure patient samples – we do not know the extent to which original anxiety leads to heart failure, or exacerbations of the condition or indeed whether the physical condition brings on anxiety and exacerbates this. We do know that many somatic symptoms overlap. I cannot refer you to any published work from my PhD as these are currently in process but should hopefully be in press towards autumn time :-)

I suggest you read the paper published in BMC Public Health. 2011; 11: 552. author David Andersson free access at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3144006/. A study possible only in the socialized Swedish health system, A question still open, I am not sure that patients of the county of Östergötland react in the same way as patients from Orange County California..

Neil Krohn from TU ? Hope this finds you well.

Just published a paper in the Annals of Family Medicine looking at General Medical Conditions and antidepressant treatment response may be helpful. The findings are as expected, the greater the physical co-morbidity, the less responsive the depression is to antidepressant treatment (this does not appear to be a function of measurement).

Your classmate David

It is not in keeping with my 30 y. long experience, centered on bodily features of depression.. With a try-and-test technique similar to hypertension treatment, I reach approximatively 90% disappearance of bodily features, like slepp, fatigue, digestive, urinary troubles, alcoholism or pain.

I am not sure if this observation will be welcome here; but there seems increasing reason to believe that what the DSM called "Axis II disorders" - personality difficulties of one kind of another - have probably the greatest co-morbidity in terms of extended hospitalisations, repeat presciptions etc.

And that's not to mention those who don't engage with treatment services at all, and so end up mired in some other part of the welfare/correction net, such as prison, homelessness and A&E.

Since Personality Disorder still carries a huge amount of stigma in the diagnosis, it tends to be under-diagnosed and under-recorded. But I can get you some reasonable estimates, if it would help.

It is obvious you cannot isolate the mental health of individuals and the general health of the society. And the end of the second World War, after 5 years of privations and hard times, the conseil national de la résistance built “la sécurité sociale” which is not only a medical assurance, but has financial responsibilities for “family funds”, “pension funds” and “working place health”.

France is not a paradise, but on the concept level, I believe UK makes the wrong choice with “National Health Service” which builds a wall between Health and Social protection. However, I must recognize the proximity solidarities are better in UK than in France probably because in this field, Britons expect nothing from the crown!

I am inclined to agree with Jacques that the creation of the NHS drove a wedge between health provision and "social provision", which successive UK governments, both Labour and Conservative, have been trying to find way to overcome for the past 20 years or more. (One UK Dept of Health minister even used the same metaphor, and talked of their being a "Berlin Wall" between health and local authoritries. )

But the "social provision" here is not just social security, meaning welfare benefits, but all the activities of the local authority, and also social housing, homelessness etc. The recent decision to return responsibility for public health back to local authorities is one successful move, at least. Other efforts seem to be always - to change metaphors - swimming against the tide.

I should really like to know more about how these things are managed in France. But this is a social policy debate; we have moved somewhat off the original subject. Still, if you (and any others?) would like to continue this, either here or in one of the other topic areas, I would certainly follow.

I'm not sure whether anyone has pointed you in this direction yet, but I don't see it popping in up in the top few answers. It might be worth taking a look at the neuro-psychological literature on "medically unexplained symptoms" or MUS. I suspect that regarding measurements of anxiety and/or depression symptoms, particularly as co-mobidities, we still cannot properly answer the question that you pose. I suspect that the field of study of MUS may have some insights to offer, in that the interactions between brain, mind and body are still poorly understood - the way that we are hard wired to cope with illness and circumstances is certainly not well understood.

However many people are working on this area of physical and emotional interactions, so with any luck, in 25 years we will have a lot more knowledge on the topic!

Dear Karen

I am not so pessimistic.

I have the diagnostic and therapeutic solution: depression is the cause of and not associated with, just because it is a neuro-biochemical illness with frequent but not constant psychic features and by no means a merely psychic disease.

For more explanations, see the abstract of my book ("La dépression et son traitement : aspects méconnus. L'Harmattan Paris 2011) in my site http://bmaroy.free.fr chapter depressionF&E, the second part being translated in in English.

Open the eyes and you shall see...

I wholeheartedly agree with Michele - a skilled and learned answer. I would like to add that converging correlational evidence points to an increase for instance in hospital visits, secondary myocardial events, fatal myocardial events and a higher level of somatic complaints in myocardial patients, if they suffer with Depression, or Depression and Anxiety. It has also been indicated that those suffering with Depression in conjunction with Anxiety have a more significant rate of these aformentioned occurences than those with Depression and no anxiety. (correlational evidence only!) In these cases there is normally a prior history of Depression or Anxiety. However, this does not indicate a cause and effect relationship. Bearing this in mind, there are shared physiological pathways bwtween the cardiovascular system and mental processes, such as emotions and cognitions, i.e., anxiety. The relationship is therefore cyclical and complex. This also is impacted by the fact that a medical illness could trigger a susceptibility to a type of affective disorder, or an affective disorder can contribute to the processes of a medical disorder, but measuring this utilising a cause and effect methodology is nigh impossible, for many reasons. With that in mind It has been found that , for instance Cerebral Ischemias can trigger what is now termed Vascular Depression, in patients with no prior history of Depression and Anxiety, or known susceptibility. In this case the Ischemia, may now be contributing to the existence and severity of Vascular Depression. Another case is that of patients who react with normal anxiety to the actions of a defibrillator (shock the heart when arrythmias occur, and this can occur anywhere, for instance in the bath) could develop avoidance to for instance bathing and over time develop anxiety disorders. Could the increase in anxiety therefore contribute to an increase in arrythmia? This type of question would require an in-depth knowledge of each patient's history, the types of medications they are taking, their quality of life outside of hospital, the type of medical disorder, the type of affective dysfunctionality and many other factors.

There are high levels of co morbidity between chronic physical conditions such as diabetes and depression and the latter is very often co morbid with anxiety. This co morbidity adversely influences the long term outcomes of both the physical and mental conditions. There is good evidence of increases in the proinflammatory cytokines in depression (which affect brain neurotransmission (5-HT and NE) and activate the stress axis (HPA). These abnormalities may be relevant to the common co morbidity between chronic physical illness and depression.

Eleni,

Would it be possible for you to provide a couple of citations on the proinflamamatory cytokines (PC) you mention? Are PCs a result of depression, anxiety or other medical conditions?

It is not clear whether PCs are a result or a cause or both of depression but immunological mechanisms are implicated in depression. I have attached a review I published last year (see section Immunological mechanisms) where this subject is very briefly discussed and the reference I quoted by Pace et al, 2007 would be useful to look at.

The interactions between “major depressive disorder” (MDD) and immunological system is a very active field of research. Hippocampal changes are observed in MDD. The level of glucocorticoids in blood is increased and glucocorticoid receptor (GR) functioning seems impaired.

Inflammatory cytokines and C-reactive protein (CRP) blood level is also increased.

You’ll find a very complete bibliography in very recent publication of T Frod and All http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3309536/

George Engel wrestled this question throughout his career as a psychiatrist when attempting to treat physical illness related to grief experiences. Perhaps to assist you in finding your scientific "center" you might read some of his work such as "the need for a new medical model" - interesting stuff.

This is a table which might help. (Ref:Medical disorders in people with recurrent depression Anne Farmer The British Journal of Psychiatry (2008) 192, 351–355)

Very interesting paper free full text at: http://bjp.rcpsych.org/content/192/5/351.long

The conclusions are specific morbidities are link to depression status, in significance order: Asthma, Osteoarthritis, Thyroid disease, Hypertension …

This study does not answer the famous question: What came first the chicken or the egg?

Both!

Mon Cher Jacques

My 30-year-long clinical experience taught me that, with an efficient treatment, associated signs disappear in parallel with classical depressive sign and reccur when stopping the antidepressant.

In think it favors clearly the responsibility of depression as the cause of the associated manifestations.

The review on CMV has now been published. There is a good section on neurological effects which will spill over into mental health issues.

http://www.hcaf.biz/2013/CMV_Read.pdf

Sorry if this offends but surely depression and anxiety are both symptoms of poor physical health issues? We acknowledge that a sick dog will mope about with a drooping tail and be prone to snapping so why not people? This renaming of symptoms into disorders and presenting them as causes for insurance and 'treatment' reasons appears to be misleading the public and it seems the professionals as well....

Well, stating the obvious, but physical dysfunction/disorder/disease can result in psychological/mental/behavioral health dysfunction/symptoms, and vice versa, hence, a positive feedback loop results regardless of original (precipitating) event/condition. This two-way confounding situation would seem to make it rather challenging to conduct clinical research into the interactions between these two domains. But thank goodness that is not my field of study!

Again the question would be... Can depression stand all alone as a cause of itself? Without stress.... social, situational, financial, physical (ie parasitical, bacterial, viral, acidic, disease etc), relational, nutritional, emotional (justified loss, fear, anger etc), environmental (heavy metals, chemicals etc) cause?

Heather, I wonder what you mean by "stand alone." All disease is to one extent or another situational and socially defined. If you are expecting to live to be age 40, then sickle cell is a protection against malaria and a heart attack is dying of natural causes. If you are expecting to live to be 70 in a malaria-free environment, then sickle cell is a disease, and heart disease becomes chronic illness if it doesn't kill you. These are both tied to genetic preconditions with the heavy influence of environmental factors. Depression is no different. There are genetic predispositions influenced by situational/environmental factors. A broader question is whether we should treat all disease in a more multi-modal fashion. Perhaps heart disease requires more than taking statin drugs and beta blockers when diet, exercise, air pollution, abdominal fat, smoking,intestinal bacteria, and oral hygiene are having an impact. Our research eliminates as many variables as possible so that we can calculate the relative impact of each of these factors alone, but in tandem their impact, as those of any variables, may be multiplicative not additive.

Hi All - Heather, I believe Sam is saying it well. Virtually no disease is "stand alone". The work of George Engel and Blair Justice (among many others) showed the interrelated nature of disease causes and effects very well, and helped distinguish "causality" from simple "correlation". Likely the best we can do as Sam mentions above is to try and isolate the relative contribution of environmental, biological, psychological, physical, behavioral, social, etc. influences on any given state of dis-ease. I agree with all above suggesting multicausal models and refer any with questions or interest to Parascandola's excellent (2001) article "Causation in Epidemiology" and Judea Pearl's book, "Causality" (2000).