Original Article					The Hong Kong Practitioner VOLUME 24 / February 2002
Patient characteristics of encounters in general practice Y T Wun 溫煜讚， T W Wong 黃子惠， W Tam 談維新， T S Yu 余德新

Summary Objective: To survey patients' demographic and socioeconomic characteristics and their morbidities (especially the respiratory and cardiovascular diseases) in general practice. Design: One-year survey. Subjects: Patient encounters from a convenience sample of 15 general practitioners (GPs, majority in the private sector) during part or the whole period of year 2000. Main outcome measure: Patients' characteristics and four categories of morbidity (respiratory, cardiovascular, other diseases, non-disease consultations). Results: The 118420 encounters were mainly from females (57%), those with full-time employment (52%), those living in the same district as the clinic (83%), and for acute diseases. About 17% of encounters by full-time workers resulted in sickness certificates. Consultations for respiratory diseases were more likely to be from current smokers. Consultations for conditions other than respiratory or cardiovascular were more likely to be from females. The respiratory and cardiovascular morbidities could be exclusively listed into a number of disease entities. Of all the encounters, 78% were for new episodes of illness and 47% for upper respiratory tract infection. Conclusion: The female gender, full-time employment, physical closeness to the clinic and acute illness were the main factors associated with GP consultations in the private sector. Keywords: General practice, morbidity, socioeconomic factors

摘要 目的：統計全科醫療中病人的人口結構、社會經濟特 徵以及疾病類別。 設計：為期一年的統計。對象：十五個醫生（主要私人執業的全科醫生） 2000年全年或某段時間開診的病人。 測量內容：病人的特徵和四類疾病人，（呼吸系統 病、心臟病，其他疾病和非病性會診）。 結果：共有118,420 位病人就診，大部份急性疾病者，女性佔57%，有全職工作者佔 52%，居住於診所同一地區者佔83%。約17%全職工作的病人需要病 假。患呼吸道疾病的病人多數吸煙。呼吸道疾病和心臟病分別分類列出，非此兩類病症者多數為女性。總 體上78%為新症，47%是因上呼吸道感染而求醫。 結論：女性、全職工作者，診所就近病人家居以及急 性疾病是私家全科醫生會診相關的主要原因。 主要詞彙：全科醫生 ，疾病類別，社會經濟因素 HK Pract 2002;24:59-65

Y T Wun, MBBS, MPhil, MD, FHKAM(Family Medicine) Associate Professor, T W Wong, MBBS, MSc, FFPHM, FHKAM(Community Medicine) Professor, W Tam, BSc, MPhil Computer Officer, T S Yu, MBBS, MPH, FACOEM, FHKAM(Community Medicine) Associate Professor, Department of Community and Family Medicine, The Chinese University of Hong Kong. Correspondence to : Dr Y T Wun, Department of Community and Family Medicine, The Chinese University of Hong Kong, 4/F, School of Public Health, Prince of Wales Hospital, Shatin, NT, Hong Kong.

Introduction Morbidity surveys have been periodically done in general practice not only to understand the diseases managed by general practitioners (GPs) but also for epidemiological studies.1-3 The last morbidity survey in local general practice was done in 1994,4 succeeding those published in 1981,5 19836 and 1987.7 All these surveys emphasised on the diseases or problems, with little reference to patient characteristics apart from age and sex. However, all clinicians deal also with people, and patients' demographic and socioeconomic factors affect the consultation pattern.8-12 It is time to have silhouettes of the human faces in relation to the morbidity patterns in local general practice. In the year 2000, our group studied the relationship between air pollution and morbidity in general practice. Our secondary aim was to survey the socioeconomic characteristics of patients going to primary care, to supplement previous local morbidity surveys. Such information would be useful to GPs as well as health care planners. Subjects and methods As each participating GP would record every patient encounter over a prolonged period, we selected GPs who were willing to commit their effort and time. Towards the end of 1999, we invited one GP from each of 11 districts in Hong Kong for a pilot study over three months. These GPs were known to be interested in research, or had past collaboration with us. After the pilot study, modification on data collection was made and part of the GP sample changed. The project continued throughout the year 2000. Data collected during the pilot period were included in the final analysis. We categorised the morbidity into four broad categories of encounters: respiratory diseases, cardiovascular diseases, other diseases and encounters not for disease management ("no disease applicable", e.g. check-ups, immunisation, cervical smears, administrative activities). We emphasised on respiratory and cardiovascular diseases as our primary aim was to study the effect of air pollution, and listed the following for detailed data collection: upper respiratory tract infection (URTI), lower respiratory tract infection, influenza, asthma, allergic rhinitis, chronic obstructive pulmonary disease (COPD), cough, hypertension, angina, and arrhythmia. Those respiratory or cardiovascular encounters that fell outside this list were classified as "other respiratory diseases" or "other cardiovascular diseases". The personal data of the patient in each encounter (one patient might have more than one encounter) included: age, sex, whether a current smoker (at least one cigarette a day), working status, living or working in the same district as the general practice. Data sheets were faxed back to the project coordinator each week or as agreed with individual GP. The GP was contacted if no data sheets were received after a delay of one week, unless prior notice was given. During the pilot period, random dates were sent to each GP for the daily numbers of patient encounters in those days and cross-checked with the entries in the data sheets. Two clerical assistants entered separately the data into electronic spreadsheets and the entered data were cross-checked for consistency. Descriptive statistics were used to describe the encounter characteristics. Cross-tabulation and multivariate logistic regression were used to assess associations, taking p<0.05 as the significant level. Result Participating GPs After the pilot period, four practices withdrew as they could not afford to meet the commitment. Five more practices from the same districts were then recruited to replace the dropouts. By July 2000, we recruited one new practice in a new town. This report excluded the data from that clinic as the patient population there had not reached a steady state by the end of year 2000. We also excluded the data from another GP who recorded only respiratory and cardiovascular encounters during the pilot period without information for other morbidities. Thus among the 17 GP practices recruited, we analysed the data from 15 (7 GPs recorded for 12 months, 5 for 9 months, and 3 for 3 months). Three GP practices were from the public sector (accounting for 10009 or 8.5% of all encounters, two participated for part of the year). The districts where the GPs practiced included Jordan, Kwai Chung, Kwun Tong, Lantau Island, Quarry Bay, Shatin, Tai Po, Tsing Yi, Wah Fu, Yau Tong and Yuen Long. Patient characteristics Age and sex There were in total 118420 encounters with some missing data for different variables; 67384 (56.9%) were from female and 50987 (43.1%) from male patients. The mean age (standard deviation) was 36.921.4 years (35.321.6 for male encounters and 38.221.1 for female encounters). Table 1 shows the age-group distribution for these encounters (data on gender or age were missing for some). The main workload (52.1%) was from adults aged 20-49 years. Our sample of patient encounters was similar to the Hong Kong Population Census 2001,13 in which the mean age was 36 years and females constituted 51.5%. However, there were more consultations from males and females from the age group of 20-29 years, and females from 0-9 years, but less consultations from males and females from the age groups of 10-19 years, and 40-49 years.

Table 1: Age group distribution of all encounters compared to Hong Kong population
Age group	Source	Male	(%)	Female	(%)	Total	(%)
0 - 9	Study HK	7883 349029	15.7% 11.0%	7037 322981	10.5% 9.6%	14920 672010	12.7% 10.3%
10 -19	Study HK	5045 439945	10.0% 13.9%	4832 415883	7.2% 12.4%	9879 855828	8.4% 13.1%
20 - 29	Study HK	8374 450765	16.6% 14.2%	12562 518228	18.8% 15.4%	20938 968993	17.9% 14.9%
30 - 39	Study HK	8837 550983	17.5% 17.4%	13264 686259	19.8% 20.4%	22106 1237242	18.9% 19.0%
40 - 49	Study HK	7310 572920	14.5% 18.1%	10608 594426	15.9% 17.7%	17919 1167346	15.3% 17.9%
50 - 59	Study HK	4820 341027	9.6% 10.7%	6482 310851	9.7% 9.3%	11304 650878	9.6% 10.0%
60 - 69	Study HK	4058 251938	8.1% 8.0%	5746 232875	8.6% 6.9%	9806 484813	8.4% 7.4%
70 - 79	Study HK	3099 159728	6.2% 5.0%	4360 183409	6.5% 5.5%	7460 343137	6.4% 5.3%
80	Study HK	935 51466	1.9% 1.6%	2005 92138	3.0% 2.7%	2941 143604	2.5% 2.2%
Total	Study HK	50361 3166801	100.1% 99.9%	66896 3357050	100.0% 99.9%	117273 6523851	100.1% 100.1%

Study HK	= =	This study, Year 2000 Hong Kong Population Census, Year 2001

Smoking status The overall rate of smoking in all the encounters was 9.1% (10780/118420). The average age for smoker-encounters was 40.3?16.4 years. With some missing data, 8502 out of 50340 (16.9%) male and 2270 out of 66875 (3.4%) female encounters were from smokers (Table 2). For encounters from patients aged between 20 and 49 years, 22.9% males and 4.6% females (12.0% in all) were from smokers.

Table 2: Age-sex distribution of current smokers
	Male current smoker					Female current smoker
Age group	Non smoker	Smoker (%)*		Sub-total		Non smoker	Smoker (%)*		Sub-total
0 - 9	7842	26	(0.3%)	7868		7005	26	(0.4%)	7031
10 -19	4749	295	(5.8%)	5044		4670	160	(3.3%)	4830
20 - 29	6268	2104	(25.1%)	8373		11677	882	(7.0%)	12559
30 - 39	6880	1955	(22.1%)	8835		12751	510	(3.8%)	13261
40 - 49	5750	1559	(21.3%)	7309		10324	283	(2.7%)	10607
50 - 59	3604	1216	(25.2%)	4820		6370	110	(1.7%)	6480
60 - 69	3237	820	(28.2%)	4057		5625	119	(2.1%)	5744
70 - 79	2670	429	(13.8%)	3099		4219	139	(3.2%)	4358
80	837	98	(10.5%)	935		1964	41	(2.0%)	2005
Total	41837	8502	(16.9%)	50340		64605	2270	(3.4%)	66875
* Percentage of the subtotal in each age group.

Working status The working status was available for 117181 (99.0%) encounters, of which 60495 (51.6%) were full-time, 1082 (0.9%) part-time, 14096 (12.0%) retired or unemployed, 17083 (14.6%) housewives, and 24425 (20.8%) children or students. Among the age group between 20-49 years, 26810 out of 36410 (73.6%) females and 23003 out of 24515 (93.8%) males were working full-time. Among 45469 encounters by females aged between 20 and 40 years 38928 (85.6%) were in full-time employment, 285 (0.6%) were in part-time work, and 3975 (8.7%) were housewives. For adult females, work employment might be an important factor for consulting GPs. Out of all encounters, 11747 (9.9%) ended with sick leave certificates, mainly for those working with full-time (10404 encounters or 17.2% of 60486 full-time worker encounters). Though female full-time workers seemed to be more likely to request sick leaves (Table 3), the average days of sick leaves (total sick days / number of encounters taking sick leaves) were same for male encounters (1.44 days) and female encounters (1.40 days). Location of GP practice Most of the encounters were from patients with their homes in the same district (98389 or 83.0%). Among those with full-time employment, 45777 out of 60490 (76%) encounters were from patients living in the same district. Females were less likely than males to be living in the same district as their GPs' practices, but more likely to be working in the same district (Table 4). Apparently, females would more likely go to GPs close to their working places but males to GPs practicing close to their homes.

Table 3: Frequency of full-time workers receiving sickness certificates Certificates given Yes No Male encounters 4751 5653 Female encounters 25074 25008 OR 0.84 (0.80, 0.87) OR = odds ratio (95% confidence intervals) (In the above table, the odds ratio of male encounters with sick leave was 0.84 compared to female encounters with sick leave)

Table 4: Distribution of patient encounters living or working in the same district as the practice Lived in same district Worked in same district Yes No Yes No Male 42875 8101 17440 33521 Female 55453 11921 25361 41958 OR 1.14 (1.10, 1017) 0.86 (0.84, 0.88) OR = odds ratio (95% confidence intervals)

Encounter characteristics and pattern From 114435 encounters, 89268 (78.0%) were new encounters while 25167 (22.0%) were follow-ups, suggesting that acute illnesses were the majority of morbidities seen in this group of GP practices. Among all the encounters, respiratory diseases constituted 57.0% (67468 encounters), cardiovascular diseases 5.2% (6205) and other diseases 34.8% (41186); only 2.9% (3434) encounters were for purposes other than disease diagnosis or management. URTI was the predominant disease for consultation (46.9% or 55582 encounters) followed by hypertension (4.2%, 5026), influenza (2.7%, 3207), lower respiratory tract infection (2.1%, 2434), "cough" (2.1%, 2546), allergic rhinitis (1.6%, 1839), asthma (1.0%, 1207), COPD (0.6%, 682), and angina (0.5%, 557). Hypertension constituted 83% (5026/6205) of all cardiovascular encounters. The lists of individual respiratory and cardiovascular diseases used in this study were proved to be inclusive. Only 1531 (1.3%) and 509 (0.4%) encounters were diagnosed as "other respiratory diseases" and "other cardiovascular diseases" respectively. As expected, significantly more encounters for respiratory diseases were from smokers (Table 5). Encounters for any other diseases were significantly less from smokers. The average ages (in years) of encounters for respiratory diseases, cardiovascular diseases, other diseases and no disease were 32.321.0, 57.816.6, 28.04.2, and 36.918.8 years respectively, showing that encounters for cardiovascular diseases were from much older aged group, and hence less likely to have full-time employment. Full-time workers were otherwise more likely to attend the GPs in this study. Logistic regression for each of the disease categories against the three variables in Table 5 and controlled for age showed that gender, full-time work, and current smoking were independently associated factors, with the single exception that full-time work was not independently associated with respiratory disease (p=0.094, exponential B=1.021).

Table 5: Frequencies of encounters by categories of diseases
		Gender		Full time worker		Current smoker
		Male	Female	Yes	No	Yes	No
Respiratory disease	Yes No	30585 20402	36870 30514	35318 25177	32066 25722	6669 4111	60772 46821
	OR	1.24	(1.21, 1.27)	1.13	(1.10, 1.15)	1.25	(1.20, 1.30)
Cardiovascular disease	Yes No	8688 48481	2506 63696	1699 58796	4502 53286	470 10310	5733 101860
	OR	4.55	(4.35, 4.77)	0.34	(0.32, 0.36)	0.81	(0.74, 0.89)
Other diseases	Yes No	16839 34111	24328 43005	21561 38901	19589 38145	3470 7304	37713 69798
	OR	0.87	(0.85, 0.89)	1.08	(1.05, 1.17)	0.88	(0.84, 0.92)
No disease	Yes No	1312 49668	2172 65189	1902 58580	1581 56190	175 10602	3305 104261
	OR	0.79	(0.74, 0.85)	1.15	(1.08, 1.23)	0.52	(0.45, 0.61)
OR = odds ratio (95% confidence intervals)

Discussions Patient encounters in this survey of local GPs consisted mainly of females (57%), those living at the same district as the practice (83%), and those with full time employment (93.8% of males and 73.6% of females of the working age 20-49 years). The most common disease was still from the respiratory system (57%), especially upper respiratory tract infection (47%). GPs in our survey did not treat many cardiovascular diseases (5%), even hypertension (4%). Encounters for solely preventive care were rare (less than 3% at the most). Among encounters from full-time workers, 17% would receive sick leaves. As in the morbidity surveys previously done in Hong Kong, the participating GPs in this study were from a convenience sample. This is a limitation from the local health care system in which centralised data from general practice (especially from the private sector) are not available and cannot be compulsory. There are often doubts on the representation of the sample and hence the external validity of data so collected. For this survey, the age distribution is close to that from the most recent Hong Kong census (Table 1). Compared to the general population, this study recruited more patients from the age groups of 0-9, 20-29 years, and less from 10-19, 40-49 years. It appears that young adults in their 20s (probably with full-time work) were more likely to seek medical treatment from the GPs in this study who were mainly practicing in private sector. This observation needs further research for its reliability and implication. This survey suggests that full-time employment, current smoking, and living close to GP practice are the socioeconomic determinants of consulting a GP. The association between full-time work and consulting a GP could be better assessed among females, as up to 36% of females aged between 18-64 years could be housewives in a local household telephone survey in 1999.14 Depending on the age group, 74%-86% of encounters from adult females in the present study were in full-time work. This is in contrast to a study of the National Health primary care in UK where the unemployed had higher consultation rates.8 There was no reason to speculate that housewives in Hong Kong were much less likely to contract any disease. Request for sickness certificates was not the major reason for consulting GP, as only 17% of all encounters from full-time workers resulted with certificates. We speculate that patients with full-time employment consulted GPs, even for self-limiting illness like URTI, probably for a quick relief from symptoms, or they were covered by insurance. Closeness between home and the GP practice is another important characteristic of attendance suggesting that patients preferred seeing GPs near their homes. A possible bias for this observation could be due to our selection of GPs in residential rather than commercial areas. GPs hence should be familiar with the paramedical and social resources in the communities of their practice. It is interesting to note that working females were more likely than males to visit GPs near their working place (Table 4). The implication is uncertain. GPs' workload was largely for respiratory diseases, especially for URTI. This survey showed that nearly 99% of respiratory and cardiovascular conditions seen in local general practice could be included into a short list. (This short list also explained 62.2% of all encounters.) Only 3% of the encounters were for non-disease (mainly preventive) care. Females and people in their 40s were the majority making use of the GPs for non-disease management. That the public does not use GPs' service for preventive care may be due to other available resources, e.g. the Family Planning Association. On the other hand, GPs advocate opportunistic and case-finding preventive care, integrating prevention into the usual consultation rather than segregating sessions for sole preventive care. The relatively few consultations for non-disease management may be an under-estimate of the GPs' preventive care. But we cannot exclude the probability that preventive care by GPs has not been widely promoted or realised. Follow up encounters were not usual, accounting for about 20% of all, and were mainly for cardiovascular diseases and least for respiratory disease (67% of the cardiovascular encounters were for follow ups while only 12% of URTI encounters were follow ups). As patients attended GPs in the same district as their home and also for acute diseases, GPs could be a good resource for taking care of their chronic diseases and related disability. The management of chronic diseases in primary care needs very much promotion. Apart from the general limitation of local morbidity studies, our survey has a few minor limitations. Half of the recruited GPs recorded the encounters throughout the whole year 2000; the other half constituted whole-year data for the districts only when combined. We do not attempt to analyse any difference between the public and private sectors, as the study was not designed for this purpose and encounters in the public sector accounted for 8.5% of all. Furthermore, two out of the three public clinics collected data only during the pilot study. Our data reflect mainly the workload as seen in the private sector of local general practice.

Key messages The patients attending private general practice had an average age of 37 years and were mainly females, similar to the general population in Hong Kong. Nine percent of all encounters (12% from those aged 20-49 years) were smokers, mostly for respiratory diseases. Full-time employment was an important factor for consulting a private GP. Chronic respiratory and cardiovascular diseases constituted less than 10% of GP encounters.

Conclusion Patients consulted GPs close to their home and mostly for acute respiratory problems. Patients with full time employment are more ready to seek medical treatment probably for quick symptomatic relief. Chronic diseases and consultations solely for preventive care are not common among encounters in private general practice. Means to promote management of chronic disease and preventive care in general practice are required. Financial disclosure The survey on air pollution and general practice morbidity is commissioned by the Environmental Protection Department to a joint group in which the authors are members.

References O'Toole B, Driver B, Britt H, et al. Using general practitioners to measure community morbidity. Int J Epidemiology 1991;20(4):1125-1132. Driver B, Britt H, O'Toole B, et al. How representative are patients in general practice morbidity surveys? Fam Pract 1991;8(3):261-268. Hansell A, Hollowell J, Nichols T, et al. Use of the General Practice Research Database (GPRD) for respiratory epidemiology: a comparison with the 4th Morbidity Survey in General Practice (MSGP4). Thorax 1999;54(5):413-419. Lee A, Chan K, Wun Y, et al. A morbidity survey in Hong Kong 1994. HK Pract 1995;17(6):246-255. Epidemiological Committee. The Hong Kong morbidity survey. HK Pract 1981;3(7):225-240. HKCGP Epidemiological Committee. Morbidity in general practice 1981-1982. HK Pract 1983;5(6):584-594. Lam P. Morbidity in Hong Kong. HK Pract 1987;9(9):2652-2666. Carr-Hill R, Rice N, Roland M. Socioeconomic determinants of rates of consultation in general practice based on fourth national morbidity survey of general practices. BMJ 1996;312(7037):1012-1013. Richards H, McConnachie A, Morrison C, et al. Social and gender variation in the prevalence, presentation and general practitioner provisional diagnosis of chest pain. J Epidemiol Community Health 2000;54:714-718. Saxena S, Majeed A, Jones M. Socioeconomic differences in childhood consultation rates in general practice in England and Wales: prospective cohort study. BMJ 1999;318:642-646. Reijneveld S, Verheij R, de Bakker D. The impact of area deprivation on differences in health: does the choice of the geographical classification matter? J Epidemiol Community Health 2000;54:306-313. Ferrie J, Martikainen P, Shipley M, et al. Employment status and health after privatisation in white collar civil servants: prospective cohort study. BMJ 2001;322:647-651. Hong Kong Census and Statistics Department. Population Census 2001. http://www.info.gov.hk/censtatd October 2001. Lam T, Chan B, Ho S, et al. Report on Healthy Living Survey: Department of Health, Hong Kong SAR, 1999.