It would be helpful to have a definition and examples of "social determinants of health."

WHO definition: The social determinants of health are the conditions in which people are born, grow, live, work and age. The social determinants of health are mostly responsible for health inequities..

A few to get started....
1. Cognitive ability of the patient. Can they remember to check their glucose at the right times instructed? Can they dose themselves for insulin?

2. Do they live alone or do they have other family or caregivers who can assist them in remembering, or performing glucose testing if they are unable to alone?

3. Are they able to maintain adequate nutrition status? Can they cook themselves diabetic diets or have a caregiver who can cook for them? Or do they need programs via insurance or community programs like diabetic prepared meals or Meals on Wheels?

4. Are they able to drive or have adequate transportation to clinical visits, including eye and foot checks, nutrition or patient education/counseling, etc. for treating all aspects of diabetes? Do they have transportation for laboratory collection for A1C, glucose (fasting glucose) levels? Or do they need home health or another entity to come and collect specimens for testing? Do they have transportation to get medication or food if needed? DO they have transportation for for urgent care to prevent unnecessary ED visits well ahead of hypo/hyper glycemic episodes before it worsens to DKA?

5. Are they able to refill and pick up their prescriptions accordingly especially when medications are changed? Do they need a home delivery service if transportation is an issue? Are they able to pay copays for prescriptions or not and it's a barrier for them to comply with their medication regimen?

6. What's their level of compliance/care for themselves? Are they always on top of everything or not, waiting until episodes occur?

7. What's their financial status and insurance status? Will they forgo treatment, care, preventative visits, meds, labs, supplies, etc., due to costs or lack of insurance?

8. Are they able to obtain and maintain their regular supplies of test strips, lancets, needles, etc.? Are there issues in them using their meters or supplied correctly?

To answer the question on usage, I've seen some of the questions asked by different health team members. However, not sure with EHRs if there is a central way of collecting the data discretely in a standardized way that is sharable across all provider roles from a communication and interoperability perspective. Perhaps can chime in on how it's collected in their different information systems?

Suspect it's mostly buried in clinical notes in a non discrete, non structured format. Some items may be in questionnaires or surveys either filled out by the patient prior to visits, or by health team members as part of their assessments.

If a SMBG lab test is ordered, the requisition would require the patient value to be provided to the performing laboratory as an Ask at Order Entry (AOE) question/field to be sent with the order and specimen to the laboratory. The patient result is often reported back with the laboratory result as part of the panel in the laboratory information system (LIS), which is then released to the EHR and other downstream information systems (registries, data warehouses, etc.) depending on configurations.

I haven't heard of a standardised collection / analyses either.

Agree, the cognitive abilities of the individual are crucial.

I wonder about the impact of poverty on ability to access test strips etc. Whilst provided free of charge in the UK, people still have to be able to get to the pharmacy regularly to fill their prescriptions.

UK medical electronic health databases already include social determinants of health. So, for example, risk calculators embedded in medical record systems incorporate measures (for example) of deprivation based on area code and learning difficulties in calculating an individual's risk of developing diabetes (bmj.com/content/359/bmj.j5019). So data is available and already being incorporated into clinical tools.

There is quite a bit on the way that deprivation leads to adverse outcomes, and I guess that use of SMBG is one of many behaviours that are affected by more fundamental problems such as depression and avoidance of self-care (e.g. drc.bmj.com/content/4/1/e00018...)

Thanks Andrew, that's very helpful

While there is an obvious need for accurate glucose data among patients treated with insulin, I think that the value of SMBG among non-insulin-treated patients is somewhat unappreciated. In my view, much of the controversy stems from inappropriate use of SMBG by many clinicians, who recommend testing but provide neither rationale for testing nor guidance in interpreting and acting upon the glucose data. Unless patients have a clear understanding of why they are testing and what they need to do with their test results, it is a waste of time and money.

This is the key question. For example, this point explains why even suggested by almost all the existing guidelines, postprandial hyperglycemia is still measured and targeted only sometime.

In the US, lack of time, reimbursement and knowledge have clearly come together to form the "perfect storm" for poor clinical management of diabetes.

Lack of knowledge by whom? When we wrote the Health Technology Appraisal Report a few years ago, it was clear that healthcare professionals were not always providing their patients with the knowledge, training or support they required in order to effectively use their meters or interpret / act on the results

Lack of knowledge applies to both clinicians and patients. However, I believe that addressing clinician lack of knowledge the other obstacles (time constraints, reimbursement issues, etc.) must be the starting point for two reasons: 1) if clinicians don't know how to provide meaningful education/training to patients, self-management will be suboptimal; and 2) even if patients become "self-educated" in their management, they will receive the support they need if their clinician remains ignorant regarding use of SMBG.

In my opinion Chris underlined the key issues. There is no doubt that also type 2 diabetes not on insulin can take an advantage by structured SMBG. This was clearly demonstrated in a controlled trial that we published some years ago:
Intensive structured self-monitoring of blood glucose and glycemic control in noninsulin-treated type 2 diabetes: the PRISMA randomized trial.
Bosi E, Scavini M, Ceriello A, Cucinotta D, Tiengo A, Marino R, Bonizzoni E, Giorgino F; PRISMA Study Group.
Diabetes Care. 2013 Oct;36(10):2887-94.

Kath, that’s a really important question. I guess the pattern of readings needed to provide structured SMBG will depend on the type of diabetes and the purpose of the monitoring. For example, in gestational diabetes, the use of paired readings before and after food can give a really helpful pointer to making the best meal choices and similarly, tagging pre-meal SMBG readings can help with titrating total medication (including insulin) dose.

Pre-breakfast readings were first used over 40 years ago to monitor basal glucose levels for people with type 2 diabetes (doi.org/10.1136/bmj.2.6137.596). Regular HbA1c checks are more widely used as a convenient and less intrusive alternative to this for many people, although for some, fasting glucose measurements can give an opportunity for more rapid increments in medication to reach optimal glucose control using pharmacological treatments. Paired blood glucose readings (night and pre-breakfast) for people with type and type 2 diabetes treated with long-acting insulin also give very helpful information about the optimal insulin dose.

It would be great to have other examples.

Andrew, I absolutely agree. Both the structured regimen and frequency of obtaining glucose patterns (e.g., weekly, once-monthly, etc.) should be considered when recommending to patients.

As far as defining structured testing vs. unstructured, I would have to say that structured testing is simply "testing with a purpose".

I also agree with Andrew's comment regarding how it depends on the type of patient, specifically for stable patients with type 2 diabetes not on insulin; there may be little advantage with home blood glucose monitoring for this group.

I agree on what all said until now. However, the concept of structured testing, as Christofer has underlined, means that the measurement must be finalized to an action. This was the confusion in the past: simply measuring does not improve the glycemic control....

Absolutely Antonio, One of the benefits of focussing on (and naming) structured SMBG testing is the opportunity to link systematic measurement to proposed actions: whether those actions are, for example, titration of medications or supporting behaviour change. Systematically collected measurements can be processed in ways (for example algorithm embedded on blood glucose meters to guide insulin dose) that make actions easier (or at least make them less difficult).

I cannot speak for anyone else, but my type 2 diabetes has been in control for years with a daily pre-breakfast glucose reading that informs me as to my overall progress. A typical glucose level under 120, with occasional fluctuations very rarely above 140, has been adequate to keep A1c under 7. This is coupled with carb intake awareness, 5 days/wk 45 min cardio exercise, metformin, stress management, and annual to bi-annual A1c testing.

Typically structured data is defined as data in discrete format which is codified with a standardized code system like LOINC, CPT or ICD-10. However, is that the intent of the question?

Is the question asking which SMBG PGHD should be collected to be clinically useful instead? Previous responses appear to address this question. As a laboratory professional, I'd add that SMBG results comparing the PGHD on their meter with the gold standard laboratory testing should be performed regularly to address whether there are precision or accuracy aspects with patient testing. SMBG results should not be comingled in the information systems with laboratory performed results, an aspect that is addressed with laboratory accreditation requirements that it is up to the laboratory director.

As one who monitors a family member's glycemic control (or lack there of at times), the means to collect SMBG results will depend on several factors including patient cognition. It may be the best tool for the patient is a paper form. While unstructured, it may be easily shared electronically via a cell phone photo with caretakers (lay) and brought into a physician office to be scanned or reviewed by providers. Not an ideal process, but may be better than nothing at all, especially for patients that may not be able to use other electronic means of discrete data collection.

On the other end of the spectrum, it would be ideal to have electronic devices collecting glucose levels (and perhaps insulin for Type 1 patients), which is connected to an app on a cell or computer where the patient can review the data, add in notes such as they overate or over medicated to provide clinicians with more context as to why glucose levels are varying. Automatic collection would help prevent typographical and transcription errors.

Ideally, said collection would then be electronically sent (insecure HIPAA) compliant communication) to the provider(s) for integration into their EHR.

The SMBG results integrated into the clinical workflow could have clinical decision support applied. CDS may include trends of the patient (downward or upward) over time. Connection with other PGHD such as weight gain or loss which may require insulin adjustments. Ideally, said monitoring may be regular (daily or weekly) for patients for low glycemic control, but may be less frequent for patients who have their diabetes under control.

Would also like to see this PGHD correlated with A1C measurements. If the patient is in control, but A1C indicates they are not, how is that reconciled by their provider? CDS may improve with the patient's data over time and more subtle changes may be indicative as predictor of illness or other issues, which may alert providers well in advance of a patient requiring an acute episode visit.

Actions by the physician based upon glucose results may be medication adjustments, scheduling additional visits, or other information communicated back to the patient. It would be nice if these changes were electronically received by the patient and marked on their system collecting their levels of when they implemented such adjustments. Then indicated on subsequent communications back to the provider to have a feedback loop, so they can see how effective the intervention was on patient care and if further adjustments are needed.

Said information should be available to all care team members whether diabetic nutritionists, endocrinologists, ED/EMS personnel handling acute episodes, etc. so they can optimize their care for the patient and provide more precision medicine.

Andrew, these are all very good points and I agree wholeheartedly. I wonder if the drive for ever-increasing technological and 'whiz-bang' SMBG technologies perhaps detracts from the benefits of simple structured testing

Would you say that the purpose has to be agreed between HCP and person with diabetes?

Absolutely agree, this is a very good point

Would you define that as 'structured'? I think so, particularly when you consider the comments above regarding SMBG being linked to action. It strikes me your testing regimen and actions are highly effective

The question was deliberately vague with the aim of gaining greater understanding of what structured SMBG is perceived to be and whether there is a consensus. Thank you for your considered response, it is much appreciated

It's a consistent systematic process that follows a structured routine, but it's not necessarily "structed" as defined by others. BTW, if I do have a reading above the 130s, I eat fewer carbs for breakfast and test again before lunch or dinner, making carb intake adjustments accordingly.

With so many great points being raised, it is difficult to pick a starting point for my comments. Although, we seem to be in agreement regarding the need to link “actionable” data with appropriate actions, I think Andrea raised an extremely important point in terms of data accuracy – inaccurate SMBG results will likely lead to inappropriate (potentially dangerous) actions.

From previous accuracy studies by Freckmann and others, we know that the relative bias (positive and negative) varies significantly between BG meter systems. As demonstrated in a recent in silico study by Campos-Náñez and colleagues (J Diabetes Sci Technol. 2017; 11(6):1187–1195), a positive systemic bias will reduce HbA1c, but increase the number of severe hypoglycemia events, total daily insulin use and frequency of daily BG testing. The issue of accuracy is further compounded by extreme test strip lot-to-lot variability found in many SMBG systems. Incorrect testing technique and other factors, such as interferents (e.g., hematocrit levels), which are unbeknownst to the user, can also negatively impact the accuracy of SMBG results. Therefore, I whole-heartedly agree with Andrea’s earlier recommendation that periodic “checks” on each patient’s meter (via laboratory reference) and monitoring of their testing technique is, I believe, fundamental to safe and effective use of SMBG, particularly among insulin-treated patients.

All points are very important, of course. The accuracy of the measurement is mandatory for any further decision. At this point, I would like to raise two questions.
Now we have a new technology (Free Style Libre): what about its precision and what about when to check glucose, considering that glycemia can be measured almost at any moment?

Great question Antonio - what do others think?

Kudva and colleagues recently published guidance for using the FreeStyle Libre (academic.oup.com/jes/article/2...) which goes into great detail on use of trend arrows and "scanning".

Accuracy is a crucial aspect - the user must be able to trust the data provided. Bad data is worse than no data often.