Intended for healthcare professionals

General Practice

Observational study of a general practice out of hours cooperative: measures of activity

BMJ 1997; 314 doi: (Published 18 January 1997) Cite this as: BMJ 1997;314:182
  1. Chris Salisbury, senior lecturer in general practicea
  1. a Department of Primary Health Care and General Practice, Imperial College School of Medicine at St Mary's, London W2 1PG
  • Accepted 24 September 1996


Objective: To evaluate an out of hours cooperative of general practitioners compared with a deputising service.

Design: Observational study of two services in overlapping geographical areas.

Setting: A general practice cooperative in Kensington, Chelsea, and Westminster and a deputising service operating in that area and the neighbouring area of Brent and Harrow.

Subjects: All patients contacting a doctor at either service in an eight week period beginning 1 September 1995.

Main outcome measures: Patients' age and sex; rates of home visiting, telephone advice, and attendance at a primary centre; hospital admission rates; prescribing rates; times of patient calls; and response times.

Results: Data were collected on 5812 patient contacts. Doctors from the cooperative visited 32.0% (1253/3920) of patients, offered telephone advice to 57.8% (2267), and saw 7.1% (278) of patients at the primary care centre. By contrast, the deputising service visited 76.3% (1444/1892) of patients and gave telephone advice to 19.3% (365). Doctors from the cooperative prescribed drugs to fewer patients (37.6%; 1473/3915) than did the deputising service (51.7%; 941/1821) (odds ratio 0.56 (95% confidence interval 0.50 to 0.63) adjusted for age and sex) and admitted 8.7% (339/3888) of patients to hospital compared with 6.8% (128/1889) from the deputising service (odds ratio 1.30 (1.05 to 1.61) adjusted for age and sex). Response times for the deputising service were faster (median time to visit 65 minutes) than for the cooperative (median time to visit 75 minutes) but the time to first contact with a doctor was shorter for the cooperative because most people initially received telephone advice.

Conclusions: The cooperative in this study dealt with patient contacts very differently from the way the deputising service dealt with contacts, fewer patients being visited and fewer receiving prescriptions. The data presented enable other out of hours services to compare their own performance using a standard data collection and analysis program.

Key messages

  • A primary care out of hours cooperative dealt with patients' calls very differently from an established deputising service, which may have important implications for patient satisfaction and expectations

  • Randomised controlled trials are very difficult to conduct to evaluate established services but could be planned before introducing a new service

  • The impact of a high rate of out of hours telephone advice on health outcomes should be a priority for future research

  • Few patients are currently willing or able to attend an out of hours primary care centre in London

  • Standard indicators and measurement tools should be defined to evaluate different models of out of hours care, and the results of an annual evaluation of local providers should be made available to general practitioners


The system of providing primary care outside normal hours has been undergoing major reorganisation, with general practitioners in many parts of Britain establishing cooperatives. These are non-commercial organisations, led and staffed by local principals in general practice, which enable doctors to spend less time on call by working within a large rota.1 In some cases cooperatives have also opened primary care centres, to which patients can be invited instead of receiving a home visit. Cooperatives often have general practitioners available at the centre who give telephone advice to many patients who contact them.

The rate of growth of the cooperative movement has been rapid1 and is likely to accelerate further as a result of recent contractual changes and new arrangements for financial support.2 However, no evaluation of a general practice cooperative has been published. Enthusiasts for cooperatives have described the success of these ventures, claiming high standards of organisation and patient care as well as high levels of patient and subscriber satisfaction,3 but the measures used have not been validated and the results not published. By contrast, there has been fairly extensive research on deputising services, though most of the work predates the 1990 contract.4 5 6 7 In an extensive review of published work Hallam argued for caution and a period of experimentation.8 She called for evaluative studies to compare alternative patterns of service delivery before widespread adoption of any particular model, a caution which has not been heeded.

A cooperative was established in the Kensington, Chelsea, and Westminster area of London in 1994 with a facility for seeing patients at a base in St Charles's Hospital. This project set out to develop instruments to evaluate general practice cooperatives, to evaluate the new cooperative in London compared with a large commercial deputising service operating in an overlapping area, and to provide baseline data against which other services can compare their performance using the same measures. This paper describes the analysis of measures of activity.

Setting and methods

The cooperative in Kensington, Chelsea, and Westminster provides cover for 147 general practitioners caring for about 271 000 patients. Between 7 pm and midnight and during the day at weekends one of these general practitioners offers telephone advice or invites patients to attend the cooperative centre. Patients are offered a home visit only when the doctor thinks this is medically necessary. These calls are passed to another doctor, who is accompanied by a driver. One to four doctors are on duty depending on the day and time. After midnight the primary care centre is closed and one doctor offers telephone advice or visits as appropriate. The cooperative operates from 7 pm to 7 am every night and from midday on Saturday at weekends. A limited service is also available on two midweek afternoons. All doctors working for the cooperative are local general practice principals, though not all members work sessions.

Before the introduction of the cooperative most general practitioners in the area used various commercial deputising services. The deputising service which has retained most subscribers locally is Healthcall plc. This is based at Ruislip and covers a large area of west London. Only 29 doctors in Kensington, Chelsea, and Westminster now use Healthcall. It was therefore decided to compare the service provided by the cooperative with that provided by Healthcall in both Kensington, Chelsea, and Westminster and the neighbouring area of Brent and Harrow, where 118 general practitioners are Healthcall subscribers and no cooperative exists.

Healthcall offers home visits and telephone advice but does not have a local primary care centre. Telephone advice is offered until 11 pm, but only to patients who specifically request advice or accept it willingly. This policy is influenced both by the 1984 code of practice for deputising services,9 which required a visit to be made to any patient who requested one, and by a responsibility not to place the subscribing doctor at risk of an allegation of “failure to visit,” as the patient's own general practitioner is responsible for the actions of a deputy. About half of the deputies working for Healthcall are general practice principals.

The populations of Kensington, Chelsea, and Westminster and Brent and Harrow have different characteristics, which may affect the calls they make and the service they receive (table 1). In particular, the Brent and Harrow area has a younger population, a higher proportion of Asians, and less deprivation than Kensington, Chelsea, and Westminster. It is important to consider differences between the areas as well as differences between the services in interpreting the results.

Table 1

Population characteristics of study areas

View this table:

Information was collected about all calls from patients received by the cooperative or deputising service on behalf of general practitioners responsible to either Kensington, Chelsea, and Westminster or Brent and Harrow Family Health Services Authorities. The data collection period extended for eight weeks, beginning on 1 September 1995. Data collected about each call included the age and sex of the caller; date and time of the call; urgency of the call as defined by the service; whether telephone advice was given by a doctor (whether or not a visit later followed); whether the final contact was telephone advice only, attendance at the primary care centre, or a home visit; time of the contact; whether a prescription was issued; and whether the patient was referred to hospital.

Data were entered from case sheets. Both services provide telephone answering only for some doctors, and some calls are received from patients making general inquiries, such as when their surgery will reopen. These calls are handled by receptionists alone at both services. Only calls which were passed to the duty doctor were analysed. Both services transfer calls to ambulance control if the situation described to the receptionist suggests a medical emergency. These calls do not result in a doctor's case sheet and were not analysed unless they were transferred to a doctor for assessment.

By using the Epi-Info computer program a data entry screen was devised which enabled a range of providers of out of hours services to enter data in a common format. The data entry program also incorporated a random number generator to select a sample of patients to be sent patient satisfaction questionnaires for another aspect of this evaluation.

Data were entered at the offices of both the cooperative and deputising services. A one in 12 sample of calls was re-entered by a research assistant to ensure the consistency of data entry between the two sites. Data were transferred from Epi-Info to spss for Windows for statistical analysis. Χ2 Tests were carried out for categorical data and Mann-Whitney U tests for ordinal data. Logistic regression was used to adjust for the effect of the confounding variables of age and sex.


Information was collected on 3920 calls to the cooperative and 1892 calls to the deputising service. Not all records were complete, the percentages below referring to records for which data were available. Patients calling the cooperative had a different age and sex distribution from those calling the deputising service. The mean age of patients calling the cooperative was 33.2 years (95% confidence interval 32.3 to 34.1 years) and of patients calling the deputising service 30.0 years (28.7 to 31.3 years). Calls about children aged under 15 accounted for 32.4% (1260/3893) of calls to the cooperative and 41.5% (785/1892) of calls to the deputising service (Χ2=46.4, df= 1; P<0.001). A total of 62.2% (2439/3919) of calls to the cooperative were made on behalf of female patients, as were 56.5% (1069/1892) of calls to the deputising service (Χ2=17.5, df=1; P<0.001).

The two services responded to the calls in very different ways, different proportions of patients receiving visits, telephone advice, or attending the centre (Χ2=1096, df=2; P<0.001). The outcome of the call was related to the age of the patient, older patients being more likely to receive visits from both services (table 2).

Table 2

Type of response to patient call by cooperative or deputising service and by age (years). Figures are numbers (percentages) of patients

View this table:

Cooperative doctors admitted 8.7% (339/3888) of patients to hospital compared with 6.8% (128/1889) of patients contacting the deputising service. Logistic regression was carried out to allow for the different age and sex characteristics of the patients contacting the two services by using the four age groups listed in table 2 as categorical variables. The odds ratio for admission from the cooperative versus the deputising service was 1.30 (95% confidence interval 1.05 to 1.61). Patients who were visited were more likely to be admitted, and the cooperative was more likely than the deputising service to admit patients who were visited (table 3). Admission rates increased with age (table 4). Analysis of whether the time of day or night affected the proportion of patients who were admitted showed no significant differences.

Table 3

Admission rates by cooperative or deputising service for different types of contact. Figures are numbers (percentages) of patients

View this table:
Table 4

Admission rates by age group. Figures are numbers (percentages) of patients admitted

View this table:

Of the patients consulting the cooperative, 37.6% (1473/3915) received a prescription compared with 51.7% (941/1821) of those consulting the deputising service (odds ratio 0.56 (0.50 to 0.63) after logistic regression to allow for age and sex). As with admission rates, there were differences in prescription rates according to how the call was dealt with (table 5) and according to age (table 6). When only face to face consultations are considered the prescribing rates for the cooperative and deputising services were 57.0% (872/1530) v 64.9% (896/1380) (odds ratio 0.74 (0.64 to 0.86) after logistic regression). The numbers of patients receiving a prescription after a telephone consultation refer to instances of doctors telephoning local pharmacies to arrange the supply of drugs (for example, to replace mislaid regular treatment or to treat a presumed urinary tract infection), prescriptions being posted later to the pharmacist.

Table 5

Prescribing rates by cooperative and deputising service for different types of contact

View this table:
Table 6

Prescribing rates by age group

View this table:

Response times were analysed for those cases which resulted in a visit, telephone advice, or a centre attendance. The most appropriate descriptive data to summarise response times are the median and 90th centile times. Table 7 gives these data according to the type of patient contact. Overall response time for the cooperative was shorter only because more callers received telephone advice. The deputising service marked 266 (14.1%) calls as urgent compared with only 11 (0.3%) calls received by the cooperative, which did not have a well defined procedure for prioritising calls. Calls designated as urgent by the services were seen only slightly faster than routine calls. The median response time for an urgent visit was 61 minutes for the cooperative and 63 minutes for the deputising service.

Table 7

Response times by type of contact and by service

View this table:

Figure 1 shows the times at which patients called between 7 pm and 7 am, when both services were open every day. The data refer to 3425 of 5794 calls (times missing for 18 calls). Of the remaining calls, 2137 occurred during the day at weekends and 232 on weekday afternoons.

Fig 1
Fig 1

Percentage distribution of calls to cooperative and deputising services between 7 pm and 7 am


This appears to be the first published evaluation of a general practice out of hours cooperative. Such an evaluation is overdue in view of the dramatic changes which have been taking place in the provision of primary care outside normal hours. The project was an observational study of two different models of service in overlapping areas. The different nature of the patient populations was exemplified by the different age distribution of callers. There may also be differences in the way doctors use the services–for example, some deputising service users carry out their own calls before midnight and others screen calls before handing some over to the service. This was reflected in the smaller number of calls handled by the deputising service despite a similar number of subscribers to the cooperative. Comparisons between the two models of organisation should be carried out with caution. Though it would be ideal to carry out randomised controlled trials to compare different out of hours services, there are few places in which a deputising service and a cooperative operate in the same area, and reorganising the services to make such a trial possible would create an artificial situation which may not represent normal activity.

This study was carried out in a metropolitan area. The results may have limited generalisability to other areas, as London has lower rates of night visits than other parts of Britain.10 However, no cooperative is representative, as cooperatives differ considerably in their organisation and setting. This project establishes baseline information and measurement tools which other out of hours services can use to compare their performance.

The most important difference between the cooperative and deputising services was in the proportion of callers who received telephone advice instead of a visit. Part of this difference may be due to some subscribers vetting calls before passing to the deputising service those needing visits, but it is also likely to be due to differences in policy. Deputising services have had a policy of making a visit to any patient who requests one, though Healthcall instituted a telephone advice service in Ruislip in July 1995, two months before this study began. Other studies have shown very different levels of giving telephone advice to out of hours callers in different settings, including 7.6% in a study of Portsmouth deputising services,11 24% and 37% for studies in London general practices,12 13 and 44% and 59% for doctors carrying out their own out of hours work in suburban practices.14 15 The appropriate level of giving telephone advice depends on both the acceptability for those patients advised and the clinical outcome. Acceptability is addressed in a separate study as part of this evaluation, but whether patients given telephone advice have a satisfactory medical outcome is an important issue for future research.

Out of hours primary care centres

The establishment of out of hours primary care centres has been seen as a priority in the new contractual arrangements for general practitioners. However, in this study only 7.1% of the cooperative's patients were seen at the centre. The progress of existing centres should be monitored before extensive resources are committed in this way. A study of Healthcall primary care centres in different parts of Britain found that 22.4% of callers were able or willing to attend,16 and unpublished figures from several cooperatives have quoted widely varying attendance rates of up to 30%.17 Interestingly, it has proved difficult to attract patients to primary care centres when a large number of patients, particularly in London, attend accident and emergency departments with primary care problems.18 An alternative strategy is to place general practitioners in accident and emergency departments19 rather than develop new centres, but this has other disadvantages by blurring the distinction between primary and secondary care.

The admission rates for both services were similar and agreed with other reports, which quoted rates of 7-8%.11 13 15 The higher admission rate for the cooperative may be due to the greater deprivation in the population served. The difference between the cooperative and the deputising service in terms of prescribing rates was more notable. Though the prescribing rate for the deputising service of 53.9% was lower than in previous studies,5 11 it was higher than the rate for patients contacting the cooperative, even when allowing for the different proportion of patients seen in face to face consultations.

Why might these differences occur? All doctors working for a deputising service are vocationally trained and some are themselves principals. General practitioners working for a cooperative are no more likely to know the patients they advise than are deputies. There may be differences in the culture of the services, the incentives and motivation of doctors working for them, and the different background of cooperative doctors who mainly work in routine daytime surgeries.

The response times for both services were disturbingly long and slower than those in previous studies,4 5 11 which probably reflected transport difficulties in London. It is unwise to place undue emphasis on response times. Many calls in the out of hours period are for conditions which are not urgent, such as ear, chest, and urinary tract infections. Patients who consult with these conditions during the day may wait several hours to be seen. It is of more concern that both services took a considerable time to visit patients they themselves designated as urgent.

Computerised management

This project was based on analysis of information from patient call sheets. Many cooperatives and deputising services have introduced computerised call management systems which will make it easier to produce such information in future.

The evaluation of out of hours services would be facilitated if services adopted standard definitions of urgent and routine calls and consistently recorded the number of general inquiries handled by receptionists alone, reminder calls, and cases transferred to the ambulance service. At a time when many alternative models for providing out of hours care are being developed it would be valuable to define monitoring requirements which are more appropriate than those previously devised for deputising services. Health authorities should require annual reports about the performance of each out of hours service provider in their areas in relation to these indicators and should make the results available to local general practitioners.

The activity measures described here form only one aspect of the evaluation of an out of hours cooperative compared with a deputising service. Further reports will describe patient and doctor satisfaction, the quality of care recorded in notes, and the costs of providing the service.


I thank Dr Sally Hargreaves, Mr Maurice Henchey, Dr Andrew Dicker, and Dr Neil Kaiper Holmes for their help and support with this project. I also thank Professor Brian Jarman and Dr Jane Wadsworth for advice, the staff at both services for their cooperation, and my research assistant Anna Marie Hill.

Funding: This project was funded jointly by Kensington, Chelsea, and Westminster Medical Audit Advisory Group and Healthcall plc.

Conflict of interest: Both organisations funding this study had an interest in the results. However, neither had any part in the design, conduct, analysis, or presentation of this work, which were entirely my responsibility.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
View Abstract