Referrals from general practitioners to medical specialist outpatient clinics: effect of feedback and letter templates

Kenny Kung 龔敬樂, Augustine Lam 林璨, Philip K T Li 李錦滔

HK Pract 2007;29:348-356

Summary

Objective: To determine whether providing feedback and educating doctors about referral guidelines will help to improve the quality of referral letters and referral rates.

Design: Prospective study.

Subjects: Patients referred from Shatin region's general outpatient clinics (GOPCs) to medical specialist outpatient clinic at Prince of Wales Hospital between 1st October 2004 and 31st June 2005.

Main outcome measures: Number of non-routine appointments, number of letters containing 1st priority and 2nd priority (P1/P2) criteria for early appointments, number of standard components included in referral letters.

Results: The number of early appointments (within eight weeks) was significantly increased (2.18, 95% CI 1.70-2.80). Doctors who have not received their specialist qualifications were the main subgroups who showed improvement after intervention [non-fellows/non-trainees: 2.70 (1.75-4.18); FM trainees: 1.78 (1.01-3.17)]. FM fellows have the highest but non-significant baseline and post-intervention early appointment rates (26.2% and 40.7%). No significant changes were noted in the number of components included in letters after intervention. Significant increase in inclusion of P1/P2 criteria was noted overall and for certain disease subgroups (blood pressure, cardiac related conditions, gastrointestinal problems and thyroid problems). In general, letters including P1/P2 criteria would result in early appointment booking (2.03, 95% CI 1.52-2.71), although this was not the case for certain disease subgroups.

Conclusion: Providing feedback and information on referral guidelines to doctors in GOPCs helps to increase the quality of referral letters.

Keywords: Referral, triage

摘要

目的： 調查給與轉介回覆和教導醫生轉介指引可否改善轉介信的質量和轉介率。

設計： 前瞻性研究。

研究對象： 2004年10月1日至2005年6月31日，由沙田區普通科門診 (GOPC)，轉介至威爾斯專科門診的病人。

主要測量內容： 非常規預約數目，轉介信符合1級和2級優先 (P1/P2) 提早預約條件的數目，轉介信中包括標準內容的數目。

結果： 提早預約(8週內)的數目顯著增加 (2.18，95%CI1.70-2.80)。未接受家庭醫生訓練的組群經上訴轉介回覆及指引後有所提高。 非院士/非受訓醫生為2.70(1.75-4.18)，經家庭醫學(FM)受訓醫生為1.78(1.01-3.17)。家庭醫學院士的轉介信得分最高，基線和介入後的提早預約率分別為(26.2%和40.7%)， 但轉介信中標準成份沒有明顯分別。整體P1/P2條例則明顯地增多，一些組別如(血壓、心臟相關問題，腸胃和甲狀腺疾病)也有所增加。通常， 包含有P1/P2條件的轉介信會獲得提早預約(2.03，95%CI1.52-2.71)，但某些病例除外。

結論： 提供回覆和轉介信指引。可以提高普通科門診轉介信質素。

主要詞彙： 轉介，分流。

Introduction

Referral implies a transfer of responsibility for some aspect of patient's care.¹ Indeed, one of the principles of management in Family Medicine is appropriate referrals.² In a world of unlimited resources, patient referrals should allow the immediate transfer of responsibility to the other party, allowing early management. However, common to all current health care systems, we are facing prioritization and inevitably long waiting time within secondary care level of public institution in Hong Kong. This "wait and delay" results in patient and medical staff dissatisfaction, while at the same time, increases adverse outcomes and health care costs, as well as reduces potential income.³ In a mixed medical economy, Hong Kong residents have learnt to "doctor-shop" to counterbalance the cost of waiting.⁴

Waiting time for the first appointment of consultation at specialist outpatient clinics (SOPC) operated by the Hospital Authority (HA) varies from one patient to another depending on the patient's clinical conditions and on the patient load of the clinic at the time. Particular factors that are considered include patient's clinical history, the presenting symptoms, findings from physical examination and investigations.⁵ All patients are then triaged into three categories: priority one (P1) for patient with the most urgent medical needs; priority two (P2) for those with comparatively less urgent medical needs; and routine for patients whose medical needs do not appear to be urgent. The median waiting time in HA for priority one and two cases are within two and eight weeks respectively.⁵ Structured triage guidelines from various specialties are available and periodically reviewed by specialists. The triage procedure is conducted by well-trained nurses and backed up by specialists. The referral letter is a key instrument in moving patients from primary to secondary care services.⁶ Efficiency of the referral process is therefore a function of the quality of referral letters, which have been shown to be of questionable standard in some overseas studies.^7-9

Of particular importance in our locality is the lengthening of waiting time for new referrals. Referrals from general outpatient clinics (GOPCs) comprise at least 30% of medical SOPCs new case referrals. Prolonged waiting time is a direct consequence of an increased workload from excessive routine referrals. Reducing the number of these routine referrals, referring only those P1 and P2 cases, and reducing the total number of referrals should logically lighten the workload in medical SOPCs, hence reduce overall waiting time. This study aims to determine whether providing feedback and educating doctors about referral guidelines will help to improve the quality of referral letters and referral rates.

Methods

Study setting and collection of referral letters

This study was conducted in Shatin District of the New Territories East Cluster of HA. Within this district, there are one medical SOPC located in the Prince of Wales Hospital and four GOPCs (Shatin Clinic, Yuen Chau Kok Clinic, Lek Yuen Clinic, Ma On Shan Family Medicine Centre). Hard copies of all referral letters sent from the above four designated GOPCs to medical SOPC between 1^st October 2004 and 31^st January 2005 were collected for the first round of data analysis. The second round of collection took place from 1^st March 2005 to 31^st June 2005. At this stage, doctors could produce either hand-written or computer-generated letters. Letters were collected both from individual GOPCs and from medical SOPC to ensure adequate coverage.

Intervention: Feedback sessions and use of referral templates

Data from the first round of data collection were presented to all GOPC doctors working in our study locality in two separate but identical sessions. These sessions were presented by the authors. Session objectives include:

1. Review current referral statistics to medical SOPCs.
2. Review number of routine appointments made.
3. Review previous letters (made anonymous) written from GOPCs, illustrating P1, P2 and routine appointments.
4. Review of the up-to-date NTEC medical triage guidelines and their relevance to patient healthcare.
5. Introduction of referral template system.

Since all GOPC doctors are using computer system during their consultations, referral letter templates containing triage criteria of medical specialist clinic were uploaded onto the NTEC GOPC computer management system for easy retrieval. Templates were made available for a comprehensive list of problems commonly requiring referral to medical SOPCs, including hypertension, diabetes mellitus, headache, chest pain and cardiovascular disease, liver diseases, asthma, chronic obstructive airway disease, obstructive sleep apnoea, renal diseases, haematological diseases, thyroid diseases and rheumatological diseases. Criteria for each disease subgroup are listed out in point form, allowing doctors to select individual criteria if present. These templates serve to allow doctors to have easy access to triage guidelines, as well as facilitating the writing of referral letters to medical SOPCs (please see appendix for an example of the letter template). In addition, doctors were encouraged to provide computer-generated referral letters.

Outcome measures

The main outcome measures were the number of components included in referral letters and the rate of non-routine referrals.

Components included in the HA standard referral letter¹⁰ were taken as the standard components that should be included. These comprise the following:

1. Name
2. Age
3. Gender
4. Identification number
5. History of presenting complaint
6. Past medical history
7. Examination findings
8. Investigation results
9. Medications
10. Diagnosis
11. Plan of management

Reference guidelines for use by doctors working in NTEC with regard to medical specialist referrals have been set up by medical colleagues within NTEC. These guidelines are used by medical triage personnel to categorise patients into 3 major categories as stated in the introduction. For the purpose of this study, only two categories were considered: early and routine. Early cases are equivalent to P1 and P2 patients, whose first appointments are within eight weeks. Routine cases are those non-priority cases whose first appointments are beyond eight weeks. Each referral letter was also checked to see the number of letters fitting the recommended P1/P2 criteria. The training background of referring GOPC doctors (trainees and fellows in Family Medicine, other specialists with fellowship qualifications and non-fellow/non-trainees) and the percentage usage of letter templates were also included as part of our data collection.

Data analysis

The number of referral letters collected was matched with the monthly medical referral statistics obtained from each GOPC. The percentage of P1/P2 appointments and the quantity of referral letters before and after intervention were calculated, from which the relative reduction in non-routine appointments was extrapolated. Likelihood ratios with 95% confidence intervals were obtained to see whether letters fitting recommended triage criteria translated to a P1/P2 appointment booking.

Results

Reduction in overall referral rate

In total there were 618 and 423 GOPC referrals to medical SOPCs in the period 1^st October 2004 to 31^st January 2005 and 1^st March 2005 to 31^st June 2005 respectively. During this period no significant changes in manpower or clinic workload was observed. The overall referral rate dropped from 0.49% per consultation to 0.34% (P<0.05). Referral templates were used in only seven letters, with six resulting in early appointments.

Differences in improvement in different doctor subgroups

Table 1 showed the number of referrals that resulted in appointments within eight weeks of the letter date, stratified according to doctor training. In all doctor subgroups there was an increase in the percentage of early appointments after intervention, although only the results in the non-fellow/non-trainee and FM (Family Medicine) trainee subgroups were statistically significant. These results were also graphically presented in Figure 1. There were no significant differences in the percentage of early appointments between different doctor subgroups at both stages.

Lack of influence of the number of components included in letters

The distribution of the number of standard components included in referral letters was shown in Table 2. There were no significant differences in the number of components included in letters before and after intervention. No significant correlation was found between the number of components included and the likelihood of receiving P1/P2 appointments.

Criteria fitting and relationship with appointment time

Relationships between the likelihood of receiving P1/P2 appointments and letter contents fitting P1/P2 triage criteria were illustrated in Table 3. Despite fitting P1/P2 criteria, referral letters for neurological disorders (before intervention) and respiratory disorders (after intervention) did not result in a significant increase in the number of P1/P2 appointments. Significant increase in criteria fitting after intervention were noted in referrals for blood pressure problems, cardiac related disorders, gastrointestinal disorders and thyroid disorders (Figure 2).

Discussion

In this study, the provision of feedback and reminder sessions to doctors resulted in significant increases in the rate of early appointment booking. Results in the "other specialist" and "FM fellow" subgroups did not reach statistical significance, but this is likely to be a result of their relatively small numbers in GOPCs. Overseas reports have also shown similar positive response after peer review and feedback.^11,12 One major factor in this improvement is the inclusion of P1/P2 criteria in letters, allowing prompt and accurate patient triaging at medical SOPCs. It is possible that in the past some doctors used letters to obtain referrals without feeling a need to provide detailed clinical history. Presumably reminder sessions help to reinforce doctors about the important "red flags", promoting triaging at the GOPC doctor level as well as educating GOPC doctors to recognize those salient features that are deemed important by medical specialists. Writing referrals is different from taking a history. A typical history must be detailed enough to rule in or rule out essential differential diagnoses. A referral letter, however, needs to contain only those essential features that help the specialist in differentiating the patient's severity. A letter that contains standard components without due regard to those essential features will not in any way facilitate management. Indeed, this may unnecessarily prolong patient's waiting time, potentially withholding proper care.

Improvement in communication between primary care physicians and specialists is possible with the use of templates,¹³ and it was hoped that its use for common diseases would serve such purpose. Despite its easy access in the computer and its user friendliness, templates were used in less than 2% of occasions in our study, although its efficacy in obtaining earlier referrals was high (86%). It is unclear why the usage rate was so low, but differences in writing styles, unfamiliarity with template usage in the computer system, and subjective relevance of the template to the actual clinical scenario are possible reasons.

The provision of guidelines on specialist outpatient triage system is not without flaws. As can be seen in this study, letter contents that included P1/P2 criteria might not result in an earlier appointment for patients. Such guidelines are set not only with respect to clinical needs, but also in relation to health economical considerations. The time lapse between the establishment of triage criteria, its implementation in SOPCs and then its subsequent distribution to GOPCs is a major contributing factor for the mismatch in appointment booking. At the moment the channel of communication between GOPCs and SOPCs is inefficient. Further improvement in quality of referral can only occur if this issue is addressed.

It is encouraging to note that letters from doctors with sufficient training in FM were more likely to result in earlier appointments for patients before and after intervention, although both results did not reach statistical significance. The lack of significance mainly stems from the relatively small numbers of FM fellows working in GOPCs and the associated small percentage of letters written. Anyhow, it can be anticipated that larger numbers of qualified family physicians could translate into a more effective primary care environment in GOPCs, at least in terms of improving referral rates to medical SOPCs. A proficient gate-keeper in the referral system needs to take into account the three perspectives on urgency before referrals: the patient sense of urgency, the SOPC doctor's sense of urgency, and his/her own urgency in the problem. It is therefore not surprising that FM fellows had better referral rates, since their training demands skills in communication between patients and professionals.

Improvement in quality of referrals should include a reduction in the total number of referrals. Although there was an overall decrease of specialist referral after intervention, it might be due to seasonal variation only. Ongoing data collection would be necessary to eliminate that possible bias. The long term impact of a single feedback session on doctors is uncertain. Indeed, such sessions would be necessary if GOPC manpower constantly reshuffles.

Limitations

This study has only looked into referrals within the Shatin region. Further analysis involving GOPCs and SOPCs in other regions will allow data to be more representative of the overall situation in Hong Kong.

Conclusion

Providing feedback and information on referral guidelines to doctors in GOPCs help to increase the quality of referral letters in terms of rate of receiving early appointments. In addition, a small reduction in the overall referral rate was seen. In view of manpower changes and the lack of long term data on the effect of a single feedback session, regular sessions would be very useful to update GOPC doctors' knowledge on SOPC triage systems.

Key messages
In comparing general practitioners with and without higher qualification, those with higher qualification are more likely to:

1. Efficiency of the referral process is a function of the quality of referral letters.
2. Prolonged waiting time is related excessive routine referrals.
3. Providing feedback sessions to frontline doctors may help in improving referral letter writing through better identification of triage criteria.
4. Continuous review with hospital specialists is needed in order to keep triage guidelines updated, such that the referral process can become more effecient.