Screening for Depression and Other Psychological Problems in Diabetes – A Practical Guide

Parts of this book hit relatively close to home and I should probably have read something along these lines some years ago, rather than now. Anyway.

Some critical remarks first. The book is not super great and parts of it are just beyond horrible, so I don’t recommend it. I gave it two stars, but this one was closer to one star than three. I wasn’t that impressed with Juth and Munthe (see also this post), but that book handles the screening stuff much better than does this one. Most of the authors of this book seem convinced that implementing some form of screening mechanism for depression in diabetics may be a good idea, but I’m far from convinced it can actually be justified. Cost aspects are somewhat neglected in the coverage, and cost-effectiveness is a key parameter in the justification process of screening initiatives; and despite what one author would like to have us believe, there’s almost zero chance such a scheme will save money in the long run – preventative medicine almost never does (Glied & Smith included a somewhat comprehensive review of these things in their coverage) and assuming otherwise is borderline arguing in bad faith. Especially problematic in terms of those things is the fact that many authors seem to agree that a screening procedure on its own, without follow-up mechanisms in place to deal with the patients after the identitification phase, probably is not justified, whereas a scheme with such mechanisms in place may be (as they put it in the introduction: “Screening for emotional problems without a comprehensive management plan has not proven to be efficacious in reducing depression and emotional problems in people with diabetes”), they don’t really talk a great deal about how this requirement of implementing proper follow-up etc. impacts the cost-effectiveness variable. Another problem is that the literature seem to find that psychiatric interventions impact quality of life metrics a lot more than they do Hba1c (in this context you can think of the latter as a variable determining to a significant extent the likelihood of developing expensive diabetes complications in the future); some authors mention this, but they are not completely clear on how this affects the cost-benefit side of the equation. The basic idea here is that if depression leads to poorer self-care behaviours among diabetics (this is not really an assumption, it’s clear that this is the case), part of this depression-mediated behavioural change may relate to lower adherence to the treatment regimen, and if so then one might think that psychiatric interventions might improve both quality of life measures and medical adherence/glycemic control measures. As mentioned it’s not clear that there’s much of an effect on glycemic control – some studies have found statistically significant effects, but their clinical relevance are questionable. Quality of life improvements are nice, don’t get me wrong, but without associated improvements in glycemic control it gets harder to justify screening – you save a lot more money by preventing a person from going blind than you do by making the guy feel better.

Some more personal comments of a less critical nature are probably in order as well. I should note that one of the most important observations made in this book – and part of why I actually didn’t really like giving it such a low rating, because it’s a very neat insight – is that it made me aware of how I may have been thinking the wrong way about depression, depressiveness and related stuff. In the past, I’ve mostly thought about depression as a dichotomous variable; either you are suffering from (major) depression or you’re not – if you do, there are specific symptom complexes which should be expected/observed (long term sleep disturbances, -changes in appetite, and so on and so forth), and if you don’t, whatever is wrong, if anything, probably isn’t a big deal. I have been thinking this way about this stuff because that’s how the DSM-IV (and V, if I’m not mistaken) approach the topic – focused on symptoms, with specific and well-defined cut-offs. The conclusion drawn on my part was that I don’t suffer from depression, because it seemed I did not meet the criteria.

If you let go of the dichotomy and start thinking about depressiveness as a continuous variable, things change. For one thing they probably get somewhat iffier in terms of empirical stuff. Mood states can change a lot over short amounts of time, and ‘objective criteria’ like weight gain may be better than unobservable self-report measures – this is presumably all part of why current criteria are the way they are. However a potential problem is that you may miss out on a lot of relevant variation by upholding a strict dichotomy, because mood states are not distributed that way in the real world (they can take on more than two values). In some patient subpopulations upholding a strict demarcation may be a lot more problematic than in others, on account of different distributions of realized mood states within subpopulations. Diabetics are probably one of the groups where it makes a lot of sense to at least think a little about how to approach people who don’t quite make the formal cut-offs (given observations made in the psycho-oncology textbook I’m currently reading, cancer patients would be another relevant patient group – and no, these two diseases are not actually that different in terms of some of the associated emotional responses to the disease; when measuring fear of progression scores based on the Fear of Progression Questionnaire, Berg et al. (2011) e.g. found quite similar scores for diabetics and cancer patients (see Goerling, page 14)). Here are some relevant remarks from the book on this topic:

“Subclinical depression is a term used when an individual presents with depressive symptoms but does not meet the criteria for a diagnosis of clinical depression. Recent reports note that approximately one-third of people with type 1 diabetes and 37–43% of people with type 2 diabetes report symptoms of depression [56, 57]. These rates were far higher than the proportion of people who had been given an actual diagnosis of clinical depression [45] . Rather than receiving treatment for depression, however, such individuals often have to cope with their symptoms alone. The impact on family, social life, and overall quality of life remains unknown to a large extent and is an area where further research is clearly needed. […] The natural course of depression is to worsen [58]”

The group of individuals with subclinical depression is likely highly heterogenous and there are some complications when dealing with this group which matter when it comes to how to approach screening mechanisms. One problem is whether the psychological distress is directly diabetes-related or not (there are measures one can use to separate non-directly-diabetes-related psychological distress from other forms of psychological distress) – this matters because different intervention types are optimal for different patient subpopulations. Another problem is that poorly regulated diabetes may actually cause physiological symptoms which mimic symptoms of depression, and that not all available screening tools which might be applied to the patient group take this into account.

With all that out of the way, a few observations from the book:

…

“In recent years, most research studying emotional problems in people with diabetes has focused on depression or elevated depressive symptoms. This has meant that depression in diabetes is the best understood emotional problem in people with diabetes. Depression rates in people with diabetes are roughly doubled compared to the general population. A meta-analysis of 42 studies demonstrated that clinical or major depression […] occurred in 11.4% of people with diabetes, whereas the prevalence in nondiabetic people was 5% [2]. People with diabetes also reported more intense depressive symptoms, without fulfilling the criteria for clinical or major depression. Elevated depressive symptoms were reported by 31% of diabetic patients, whereas only 14% of nondiabetic subjects reported elevated depressive symptoms. The doubling of depression rates in people with diabetes compared to nondiabetic people has been confirmed by a more recent meta-analysis [35].”

“The negative impact of the comorbidity of diabetes and depression on quality of life is greater than the sum of diabetes and depression alone, indicating an exponential detrimental effect of depression on quality of life in people with diabetes. Although depression is a rather common condition in chronic diseases [47], a WHO World Health Survey on quality of life in different chronic diseases (arthritis, asthma, angina, and diabetes) showed that quality of life was most impaired in diabetic patients with depression [48].”

“In a prospective study with 7-year follow-up, Black and colleagues demonstrated that the risk for macrovascular complications was more than three times higher if depressive symptoms were present in diabetic patients at the start of the study [11]. The risk of developing microvascular complications or functional disabilities in diabetic patients with minor depression is increased by a factor of 8.6 or 6.9, respectively. Interestingly, the risk difference for late complications between those with mild and more severe depression was rather small. Thus, it seems that even milder forms of depression have to be taken seriously. […] the experience of depressive symptoms that would not meet the diagnostic threshold for MDD is a risk factor for negative health outcomes in patients living with diabetes […] data clearly demonstrate an incremental relationship between symptoms of depression and negative health outcomes in diabetes, a relationship observed even at subclinical levels of depression severity. [This] challenge[s] the model of MDD in diabetes, which conceptualizes the problem of depression as a categorical construct that is either present or not.”

“Until recently, there has been a paucity of evidence about the treatment of depression in people with diabetes, and consequently there has been uncertainty about the most effective and safe way to do so […] The effectiveness of psychological interventions in people with diabetes has [however now] been demonstrated in a systematic review of 25 randomized controlled trials of psychological therapies, mostly CBT. Both psychological distress and glycemic control were improved in people receiving active psychological interventions [60]. A further systemic review of 29 trials and meta-analysis of 21 trials by the same group showed that psychological interventions improved glycated hemoglobin by approximately 0.5% (5 mmol/mol) in children but not in adults [61]. […] recent reviews by David-Ferdon and Kaslow [94] and prior work by Kazdin and Weisz [95] highlight the following components as primary targets of CBT: (1) increase participation in pleasant activities (that enhance mood), (2) increase and improve social interactions, (3) improve conflict resolution and social problem-solving skills, (4) reduce physiological tension or excessive affective arousal, and (5) identify and modify depressive thoughts and attributions.”

“Diabetes management in older patients presents unique challenges. Clinical (e.g., comorbidity, complications) and functional (e.g., impairment, disability) heterogeneity in the older population require special attention. Most diabetes patients have at least one comorbid condition [1] and as many as 40% have three or more distinct conditions [2].”

“Diagnosis and treatment of comorbid depression in older patients is a considerable challenge in routine diabetes care. Depression is frequently under-recognized and under-treated [51–54], with less than 25% of diabetes patients’ depression successfully identified and treated in clinical practice [55].”

“The risk of incident foot ulcers has been found to be increased twofold in individuals with comorbid depression compared to diabetic patients who are not depressed [9]. Depressed patients with diabetic neuropathy are more prone to developing first foot ulcers than nondepressed individuals, independently of biological risk factors and foot care [10]. […] There is also strong evidence of an inverse association between diabetes complications and depression. Patients burdened by diabetes complications are more likely to develop depression than are those without complications, especially in the case of nephropathy and neuropathy [13]. […] Depression is common in patients with erectile dysfunction, which reflects a continuous interplay between diabetes-related and psychological factors. […] There is substantial heterogeneity between type 1 and type 2 diabetes comorbidity with depression, which is partly explained by their different etiologies [8].”

“Overall, findings derived from reviews and individual studies suggest that more research-based evidence is needed to support the case for the widespread introduction of screening for depression in people with diabetes in primary care, or indeed in other settings. A recurrent message is that screening alone is unlikely to have a strong impact on patient outcomes unless case-finding is linked to other aspects of patient management. […] it remains to be shown that formal pro-active screening has benefits over improved methods of incorporating recognition and management of depression into routine models of care of people with diabetes.”