A pilot follow-up study on the use of a reminder system among patients with unsatisfactory control of diabetes mellitus in a Hong Kong public family medicine clinic

Kar-fai Lee 李嘉輝, Man-li Chan 陳萬里, Liang Jun 梁峻

HK Pract 2011;33:56-62

Summary

Objective: A pilot follow-up study was designed to evaluate the use of a reminder system in improving the management of diabetic patients with unsatisfactory control in a primary health care setting. Such patients were defined as those with HbA1c level ≥ 9%.

Design: A cohor t study. Elect ronic reminders to improve diabetic control were entered in the computer record of patients with HbA1c ≥ 9%. The change in HbA1c was analyzed.

Subjects: All patients attending the clinic from 1^st November 2006 to 31st October 2007 with HbA1c level ≥ 9% and continued to have follow-up in the clinic in March 2008 were included. Patients who died, were referred out or defaulted appointment were excluded.

Main outcome measures: Doctors' response rate to the reminder system suggestions, change in the HbA1c level of patients at the start and end of the study.

Results: 57 patients were recruited. Doctors were active in using different tools to improve diabetic control. The response rate to reminders reached 93%. The average HbA1c was reduced by 0.76% (p=0.001) after the intervention.

Conclusion: A reminder system can be useful in improving diabetic control although further studies with more rigorous design will be needed. It seems that it is practical and well accepted by doctors. Further study on the use of reminder systems with multicenter involvement and use of control group is worthwhile.

Keywords: Diabetes mellitus, reminder system, suboptimal diabetic control, HbA1c.

摘要

目的： 以試驗性跟進式研究評估在基層醫療環境下，是否可以利用一個提示系統改善糖尿病控制欠佳病人的治療。該等病人為其HbA1c水平在9%或以上。

設計： 一項群組方式研究。在HbA1c為9%或以上糖尿病人的電腦病歷紀錄內加入提示標記，提醒醫生積極改善糖尿控制。在跟進時，對HbA1c的變化加以分折。

研究對象： 所有從06年11月至07年10月，HbA1c水平在9%或以上的糖尿病人，而在08年3月仍然回來覆診者便會被納入是項研究。身故、被轉介出或未再回來覆診者均不會被納入研究。

主要測量內容： 醫生對提示系統提議的順從性。病人的HbA1c在研究前後的變化。

結果： 57名病人被納入研究。醫生均積極地用不同方法去改善病人的糖尿病控制。對系統提示的順從達到93%。平均HbA1c因而下降0.76%（p=0.001）。

結論： 提示系統可以是一種適用方法來改善糖尿病控制，但仍需要作進一步研究確實。它似乎是可行的，而且頗受醫生們接受。因此，值得再進一步在多個中心進行集體研究，並加入病人比對組別以作比較。

主要詞彙： 糖尿病，提示系統，控制欠佳，糖化血紅素。

Introduction

Diabetes is a very common chronic disease. The worldwide prevalence was estimated to be 2.8% in 2000 and 4.4% by 2030.¹ It is higher in developed countries. The total number of people worldwide with diabetes is projected to rise from 171 million in 2000 to 366 million in 2030.¹ Diabetes causes about 5% of all deaths globally each year.² In Hong Kong, the age adjusted prevalence was 7.7% in 1990 and increased to 8.5% in 1995.³ The elderly population (more than 65 years old) has an even higher prevalence.⁴ It is the ninth commonest cause of deaths in Hong Kong.⁵

Poor diabetic control is associated with worse health outcome. It was confirmed in a large prospective cohort that increased HbA1c was an independent risk factor for cardiovascular disease. For each 1% increase in HbA1c, the hazard ratio for cardiovascular disease was 1.08.⁶ Healthcare costs also increase with increasing HbA1c levels.⁷ Improvement in diabetic control decreases disease burden at the patient and society level.

Reminder systems have been used in different ways to improve clinical care of patients. There is a wide variety of clinical reminders. Some target on doctors and some are for patients. It can be patient specific or generalized. The format may be electronic, phone-based or paper-based. Studies showed that reminder systems improved smoking-status documentation,⁸ vaccination uptake rate,^9-11 health screening ^12-14 and adherence to medication.¹⁵

For diabetes, reminder systems have been used in screening and management processes. A patient and computer reminder system in New Zealand was shown to be effective in increasing screening for diabetes. ¹⁶ Doctor-oriented HbA1c tracking tool ¹⁷ and real-time reminder protocol ¹⁸ are useful to improve management of diabetic patient and reduce complications. Patient-oriented point-of-care computerized diabetes-care reminder system was accepted by both patients and providers and improved the quality of outpatient-care for diabetic patients.¹⁹ Computerized reminder systems also help to improve compliance with recommended care by facilitating the documentation of clinical findings and the ordering of recommended procedures.²⁰ Study on the use of reminder system for clinical care of diabetes patients with unsatisfactory control have scarcely been done.

Objectives

1. To evaluate the use of a reminder system in improving diabetic management in a public family medicine clinic especially for those with poor diabetic control.
2. To evaluate doctors' response rate to the use of reminder system.
3. To evaluate the initiation of insulin with the aid of the reminder system.

Method

The study was conducted in Tuen Mun Hospital family medicine clinic in 2008. It is a public hospital in New Territories West Cluster of Hong Kong Hospital Authority. The clinic is fully computerized. There are about 600 patients with diabetes mellitus in the clinic. Diabetic complication screening is performed annually and results including HbA1c level are recorded in the computer system. The study was approved by the Tuen Mun Hospital Research Ethics Committee.

Subjects

Patients with diabetes mellitus attending the author's clinic from 1^st November 2006 to 31^st October 2007 satisfying the following inclusion criteria were included:

• HbA1c ≥ 9%

• Continued to have diabetes follow up in the clinic on 31^st March 2008

Exclusion criteria included: patients who died, were referred out or were lost to follow-up.

Study design

This was a cohort study. The latest HbA1c level within 1 year before the start of the study (March 2008) were recorded and compared with the HbA1c level within 6 months before the end of the study (June 2009).

Intervention

An electronic reminder was created in the Hospital Authority Clinical Management System [CMS] by one of the authors who is a family medicine specialist. Content of the reminder was based on the departmental management consensus in March/April 2008. It would pop up on the computer screen every time the patient's record was accessed. It would remind the attending doctor that it was a case of poor diabetic control and suggested several ways to improve glycemic control. They included 1) referring patient to community care centre or dietitian for diabetic health education, 2) referring patient to diabetic centre for initiation of insulin, 3) adjusting drug dosage (oral hypoglycemic agent or insulin) and 4) close monitoring of diabetic control by checking HbA1c half yearly. These four points were generically included in the reminder content.

Outcomes

The study lasted 1 year. Patients' records were reviewed in June 2009. The primary outcome was the change of HbA1c level. The time period between pre and post intervention HbA1c level checking was about one year. The secondary outcome was doctors' response rate to the reminder and the use of insulin. Patient demographics and duration of diabetes were also collected.

Statistical method

The change in the HbA1c profile before and after the study was tested with the paired student's T-test.

Results

57 patients were recruited. One patient died of cancer. Another patient was referred out because of dementia. 55 patients continued to have follow-up in the clinic at the end of the study.

Patients characteristics

The patient profile is shown in Figure 1. The average HbA1c at the start of the study was 10.03%.

Outcomes

Doctors were active in using different tools to improve diabetic control including referral to community resources and diabetic centre, increasing oral hypoglycemic agents or insulin dosage and more frequent checking of HbA1c level (Figure 2). The overall response rate to the reminder reached 93%. (Response was counted if one or more of the reminder recommendations was followed)

The average HbA1c was improved by 0.76% from 10.03% to 9.27% at the end of the study (p=0.001) and the 95% confidence interval of HbA1c improvement was 0.35 to 1.17. Fifteen patients (27%) newly received insulin treatment after the study. Forty patients (73%) showed improvement in HbA1c level.

Discussion

The reminder system was practical and easily adopted in the author 's clinic. It was convenient to draw a patient list with poor diabetic control in a computerized clinic with data retrieved from the Hospital Authority information technology department. It was also well accepted by doctors as reflected by the high response rate to the reminder.

The reminder response rate was high. Previous systematic review showed computer reminders improved adherence to processes of care by a median of 4.2%.²¹ The doctors were active in using different tools to improve HbA1c control. The authors proposed that the active behavioural pattern was related to the implementation of the reminder system and this in turn contributed to the improvement in HbA1c value in the reminder group. The results were encouraging.

Evidence generated from the United Kingdom Prospective Diabetes Study shows that HbA1c reduction results in significantly lower risk of developing macrovascular and microvascular complications.²² The better diabetic control as reflected from the decreased HbA1c should help to decrease health burden on patients and health care cost.

Setting up a reminder system involves identification of target patients and input of reminder. This is a time consuming process and may cause difficulties for other family medicine clinics or primary care clinics with a large number of diabetic patients. It could be facilitated with the use of computer system to identify target patients. In the setting of the Hospital Authority clinics, this can be achieved with the help from the information technology department. For clinics using computer system to manage patient record, e.g. Clinic Solution version 6.0 the software can sort out a list of patients with high HbA1c level. The input of reminder could be helped with standardized template and mobilization of other clinic staff, e.g. clerks and nurses.

Limitations

The small sample size may affect the generalization of the result to other practices. Because of the lack of the control group, the improvement of HbA1c level after the study may not be solely contributed by the reminder system.

Conclusion

Reminder system is useful to improve diabetic care for those with poor control. It is likely that this kind of system is also useful in the management of other chronic illness, e.g. hypertension and hyperlipidaemia. Family physicians may consider setting up reminder system to improve chronic illness management and it is practical and easy to set up such systems.