The many dimensions of DAWN2

Rhys Williams

I am extremely pleased that the fourth issue of Diabetes Voice this year is devoted to the DAWN2TM study. This second DAWN has built on and extended the work of the first (2001) DAWN by shedding more light on the attitudes, wishes and needs not only of people with diabetes and healthcare professionals but also of members of families in which diabetes is present.

In his editorial in the October 2008 Diabetes Voice issue devoted to the DAWN Youth study, Professor Martin Silink, then President of IDF, wrote ‘Diabetes imposes a heavy burden of care, leading not only to financial pressures, but also to stress and anxiety for all the family’. DAWN2 has significantly added to our knowledge of several facets of this family stress and anxiety. Information has been provided, by over 2,000 family member respondents recruited from the 17 DAWN2 countries, about: the burden they feel diabetes imposes on their lives; the distress caused by diabetes; the worry specifically related to the risk of their relative becoming hypoglycaemic; their frustration in not knowing ‘how best to help’ and the proportion participating in any diabetes educational programme or activities (just over 20% on average with highs for respondents based in Canada and Denmark (≈40%) and a low for those from the Russian Federation (≈10%)). This is complemented by important information provided, from these same 17 countries, by around 8,500 adults with diabetes (both type 1 and type 2) and just over 4,700 healthcare professionals responsible for the care of people with diabetes. Among the strengths of DAWN2 are the description of the many facets of diabetes from these several viewpoints, the fact that these descriptions have been constructed on the basis of well-validated research instruments – some ‘off the shelf’ and some created specifically for this study – and the availability of potentially rich free-text qualitative data giving individuals’ views and experiences.

The choice of methods in DAWN2 is a lesson in pragmatism. (For further details, see the first of the Diabetes Research and Clinical Practice articles listed a few pages further on in this Issue.) The people with diabetes and family members were recruited by a mixture of methods – ‘web, telephone and in-person’ while the healthcare professionals, similarly, were identified from a mixture of sources – ‘online panels and databases’ and ‘other sources such as telephone lists and physician directories’. The method by which responses were obtained was described as ‘hybrid’ – different in the countries regarded as having ‘high internet penetration’ from those considered not to have such high penetration. Whatever was likely to work seems to have been used.

Methodological purists may well be uncomfortable with this heterogeneity of methods. An important question to be asked is: were these samples in any way representative of the totality of people with diabetes, family members and healthcare professionals in their countries? There really is no way of telling. Did the several methods of obtaining responses affect these responses in different ways? Again, there is no way of telling. The challenges and limitations of conducting such a study in this way are reasonably fully acknowledged in the three peer-reviewed papers that have already been placed in the public domain (and quoted in several articles in this Issue) and must be borne in mind when the results are assessed. Specifically mentioned in these published accounts is that we know nothing about the characteristics of any potential respondents that might have refused to take part.

Notwithstanding the above caveats, the DAWN2 results further emphasise how far we are away from eliminating or even significantly reducing the negative psychosocial effects of diabetes on individuals and on families. We are promised, in future publications, comparisons of the views of patients, family members and healthcare professionals about health services within countries. Discordance between these views is likely and will be particularly useful for illuminating ways in which these services can be improved. As mentioned above, as well as having responses to closed questions, DAWN2 has extensive free-text answers. Rigorous qualitative analysis for common themes (for which representative samples are not necessary) will be particularly useful.

Finally, if there is to be a ‘DAWN3’ how can progress (or the lack of it) be assessed? To use the same mixed bag of sampling methods as DAWN2 just will not do. (DAWN3 samples for India, for example, may by chance have totally different characteristics from those of DAWN2 for the same country.) To re-contact the same individuals, in addition to recruiting further respondents, would be one theoretical solution. Better still would be to embed future data collection in routine health surveys, using the same or a similar set of instruments, and making use of truly representative samples of the populations of the same 17 countries (and others if at all possible). It would be a shame if the DAWN2 observations obtained with such effort were to be confined to just one moment in time.

Readers may find some repetition in the articles in this issue. We have attempted to keep this to a minimum. Personally, I find this reassuring and view it as consistency and reinforcement rather than repetition.

Editorial, type 1, type 2, diabetes, DAWN2 study