Changes in Office Visit Use Associated With Electronic Messaging and Telephone Encounters Among Patients With Diabetes in the PCMH

This article sought to examine a relationship between frequency of secure, electronic messages or telephone calls made to patients and subsequent frequency of office visits. The authors describe their adoption of a Patient-Centered Medical Home (PCMH) model as guiding their study, however, the application of PCMH was not clearly conveyed by the authors. The number of office visits was the primary outcome measured. We discussed wanting to understand the significance of the number of office visits as several questions were posed by the group such as, should we desire fewer visits and thus less cost or should we hope for closer patient follow-up including both messages/calls and in-person visits? The authors adequately provided a background of evidence regarding the relationship between messages or calls and office visits.

The group discussed the importance of office visits within the PCMH framework. We discussed the importance of measuring the effect of messages and calls on access to care, cost, and quality of life (i.e. reduced morbidity and mortality of the chronic disease state). We found that the study contributed to understanding the relationship between the use of low-cost, electronic and telephone communication and office visits.

The group thought the findings were too limited to guide clinical practice. As the authors state, there are several limitations arising from the methods of this study. Tracking only patients with diabetes limits the generalizability of this study significantly. The chronic disease management of diabetes varies significantly from that of other chronic diseases, requiring a more active role from the patient. Perhaps the effect of messages or calls would be different among the general population or patients managing other chronic diseases.

As the authors point out, the initiator of each contact was not tracked. It is conceivable that a patient who has begun to call in with greater frequency might also desire to be seen in person, thus biasing the results. Any connection between an increase in messages or calls and an increase in morbidity was not examined, but would conceivably be related to a necessary increase in office visits as well. In cases where contact was initiated by the providers, we wondered if a protocol was developed directing either the use of messaging/phone calls or their content for standardization. Ideally, a message or call would either be scheduled regularly or would be triggered by an event such as a high fasting blood glucose reading, and would deliver a consistent message across all patients. As the authors state, "secure electronic messaging and telephone encounters had been conducted at Group Health for several years," we wondered what precipitated the increased rates of messaging and calls. Without protocols governing the rationale for making contact, we cannot know. Not controlling the content of messages might also result in wide variability in patients' desire for follow-up, thus potentially skewing results.

Additionally, in our experience, there is a wide range among diabetic follow-up office visits from 15-minute check-ups to much longer visits addressing both planned interventions (lifestyle counseling, patient empowerment, self-management education) and unplanned patient concerns. We postulate that frequent messaging/phone contact might alleviate some of the extra patient concerns that can occupy more than the allotted time in an office visit. Analysis of this data set examining the time spent (and thus cost) rather than number of office visits may have provided more meaningful data. Without knowing the content or reasoning behind the increased frequency of contact, we cannot know the actual effect it may have had on office visit frequency. We would be curious to see an analysis of the raw number - rather than percent increase - of messages sent or calls made to a patient and subsequent visits (correlation between number of contacts and number of visits.

Of interest to our group were the findings presented in Table 1 with regards to stratification by age. While encounters were more frequent among older patients at baseline, we were surprised to find that patients 55 and older used secure message threads at the same rate or more frequently than their younger counterparts. This suggests that electronic messaging may have reasonable acceptability among the older patient population in this study. Further investigation of this could inform future use of healthcare messaging among older populations.

Using markers of diabetic morbidity - hemoglobin A1C, blood pressure, and LDL levels - the group assigned patients to moderate, high, and very high morbidity burden groups. We suspected - and the data appears to support that increased contact might benefit higher morbidity patients more than those with low or moderate morbidity. Indeed, data presented in Table 2 indicates that patients with very high morbidity burden had a significant decrease in office visits. We would have liked the authors to explore and discuss this finding. In the setting of a PCMH, we suspect even greater meaning in this finding given the potential for reducing costs of care in some of the more challenging and closely monitored patients.

Overall, the group thought this study addressed an increasingly important issue and did so with an impressively large data set. With more care, this study might be very informative in both the implementation of secure electronic messaging and the PCMH model. To do so, we would like to have seen more measurable outcomes, or at least clarification of the significance of increased office visits as an outcome. Looking at cost, quality of office visits, or measures of morbidity might be more informative in determining the efficacy of both secure electronic messages and telephone calls as interim interventions. In the end, we were uncertain how the study findings would significantly change our clinical practice and the management of chronic disease patients within or outside of a PCMH model.

Competing interests: None declared