A new marker of primary care utilization - annual accumulated duration of time of visits

Primary care visits

The primary care visit remains the principal opportunity for health care providers to address patient’s needs. The results of the Israel Central Bureau of Statistics (ICBS) for 2009 indicate that the annual average number of visits to the primary care physician (PCP) is 6.2 in the general population of Israel and 16.1 for ages 65 and over. Age and the number of visits of patients with chronic diseases were found to be factors that significantly increase the annual average number of visits [1]. The most recent data found by us suggests that the mean duration of a visit with an Israeli PCP is 10.4 min [2].

The annual average number of visits can vary substantially across countries. One study in the United States calculated a mean of 1.6 PCP (defined as visits to a general practitioner, family physician, pediatrician, geriatrician, or general internist) yearly visits per person as of 2008 [3]. In the WHO European Region, the average outpatient contacts per person per year in 2006 was 7.85, and country specific averages for 2006 or the latest available year were 7.0 in Germany, 9.5 in Spain, 5.4 in the United Kingdom, 5.7 in the Netherlands, 6.6 in Belgium and 11.0 in Switzerland [4].

There is also significant cross-country variation in visit duration. In the United States, 2006 data from the Centers for Disease Control and Prevention (CDC) found that the mean duration of face-to-face visits with PCPs (general or family practice) was 19.5 min [5]. In Europe, it was found that the mean length of a visit with a PCP (general practitioner) was 7.6 min in Germany, 7.8 min in Spain, 9.4 min in the United Kingdom, 10.2 min in the Netherlands, 15.0 min in Belgium and 15.6 min in Switzerland [6]. A study by Bindman et al. found in a 2001–2 cross-sectional analysis that the average duration of a face-to-face visit with a PCP in the US (general internists, general pediatricians, and family practitioners) was 16.5 min, about 10% longer than with general practitioners in Australia (14.9 min) and New Zealand (15 min). Visit lengths were longer in the US for all age and gender groups. Because the average number of primary care visits per capita was greater in New Zealand and Australia, however, the mean per capita annual exposure to primary care physicians in the US (29.7 min) was about half of that in New Zealand (55.5 min) and about a third of that in Australia (83.4 min) [7].

Studies from various countries have found that the length of an ambulatory visit with PCPs is influenced by increasing age, presence of psychosocial problems [8], gender (women) and greater number of new problems discussed in the visit [6].

Population and data source

Data was retrieved from the Clalit Health Services (CHS) central computerized database. CHS is the largest HMO in Israel, covering 54% of the entire Israeli population (about 4,200,000 people in 7 districts). Every person insured by CHS is assigned to a PCP. All the visits to a PCP are fully computerized and the information from the electronic medical records is retrieved to a central repository. The central database includes demographics, information about physician visits, and a register of a selected number of chronic diseases (from the HMO’s registry, diagnosed previously to the visits in question).

The study period was the entire 2012 calendar year. The population of this study consisted of all adult members of the HMO aged 20 and over, from which we draw a national random sample of 83,707. The sampling method was a randomized computer based binary extraction of 2% of all patient data, based on the two last digits of the patients’ social security number.

Of the patients who were randomly selected from the HMO’s database, 1088 died during the study period and 2615 left the HMO. Patients older than age 100 years (n = 25), bed-ridden (n = 2059) or in a nursing home (n = 673) were excluded from the study; therefore, the current analysis included 77,247 patients.

Data accessed

The number and duration of visits of CHS members with a PCP were retrieved for the study period. Additional patient data included: demographic characteristics: age, gender, country of birth, year of immigration to Israel (Individuals who were born in Ethiopia and immigrated to Israel after 1984 were defined as “new immigrants”. Immigrants from other countries were defined as “new immigrants” if they immigrated after 1990. These represent the two major waves of immigration to Israel that took place in the past 30 years), residency (Large city ≥100,000 citizens, other city, collective settlement - also known as a Kibbutz, cooperative Israeli settlement, small town and non-Jewish settlement), socioeconomic status (SES; low SES was defined as exemption from social security payments); chronic diseases (malignancy, diabetes, hypertension, hyperlipidemia, ischemic heart disease (IHD), chronic heart failure (CHF), status post cerebrovascular accident (s/p CVA), asthma, chronic obstructive pulmonary disease (COPD), dementia, epilepsy, anxiety disorder and drug abuse); and a Charlson comorbidity index [24, 25], which was calculated as well.

The study was approved by the CHS ethics committee at the Meir Medical Center, Kfar Saba, Israel.

Statistical analysis

Descriptive statistics was the primary method of analyzing the data. The annual number of visits and annual duration of visits (in minutes) were analyzed as continuous parameters. The Central Limit Theorem justifies the results despite the non-normal distribution of these variables.

Demographic characteristics were compared as well as medical characteristics for sub-groups according to number of visits and visit duration, using correlations (for differences between continuous parameters), T-tests (for differences between dichotomized parameters and averages of continuous parameters), chi-squared analysis and Fisher IS (for categorical parameters) and ANOVA (for differences between more than two categories in a parameter). If the ANOVA was found to be significant, a POST HOC analysis using Tukey’s test was performed to distinguish the different categories.

We used multivariate analysis to construct predictive models for comparison between annual number of visits and annual duration of visits.

A Multivariate Linear Regression model was applied to the data to study simultaneously the independent relationship between the demographic (age, gender, SES, residence area, and immigration status) and clinical background (chronic diseases, Charlson comorbidity index) and visit characteristics. The model predicts the probability of higher number of visits and longer annual duration of visits as a function of the explanatory variables. We addressed the non-normal distribution of these variables by using a square root transformation.

A p-value of 0.05 or less was considered statistically significant. All results were rounded to tenths (+1 decimal place). All analyses were carried out with the assistance of The Statistical Consulting Lab at The School of Mathematical Sciences at Tel Aviv University, using SPSS ver. 21 statistical software.

Summary

During 2012, the average annual number of visits with a PCP was 8.8 ± 9.1 and the median was 6 ± 10 IQR. The mean AADT was 65.7 ± 75.8 min and the median AADT was 43 ± 75 IQR minutes. The average duration of a single visit was 7.6 ± 4.3 min and the median was 7 ± 4.5 IQR minutes, which is lower than the data known to us prior to this study [2]. This was to be expected following the rise in the PCP’s workload due to population growth and the increase in life expectancy.

The main characteristics of patients with a higher annual number of visits and a higher AADT with a PCP were: female, older in age, a higher Charlson index (all three of which coincide with previously known data [1, 6]), of a low SES (which could be explained by Israel’s public health care system, providing highly available/no cost primary care), and residing in a kibbutz (possibly due to greater accessibility to PCP’s). New immigrants had a lower annual number of visits and a lower AADT with a PCP.

The study also found that average visit duration was very similar for both men and women and across the various age groups. This implies that the relationships of age and gender with the number of visits are similar to their relationships with AADT. Thus, while AADT does a better job of capturing resource use (i.e. the amount of time physicians invest in the care of various types of patients) than does the number of visits, it is reasonable to continue using the number of visits as a proxy for AADT in calculating capitation formulae. In the future, it will be important to examine whether visit duration is also consistent across geographic areas.

Strengths and limitations

One of the main strengths of the study is that it was based on a national sample from the largest HMO in Israel. Another is its use of thousands of electronic medical records (and not self-reports) from hundreds of general practices. This is in comparison to other studies, where the exposure to primary care was calculated from duration of visits recorded by the physician, and reports on rates of visits to primary care for each country [7, 22, 26, 27]. However, international comparisons may be affected by differences in definitions and in the circumstances in which patients see primary care physicians in different countries. It is possible that some references to outpatient attendances include in part visits with specialists.

Another issue is that there are a substantial number of physician visits that are administrative in nature (repeat prescription, fill out laboratory tests forms, etc.) and do not entail a face-to-face meeting between patient and physician. Although the type of visit is specified in the electronic file, in our experience, this information is usually not accurate and therefore the type of visit is difficult to determine. Therefore, we could not separate between face-to-face and non face-to-face visits, but we believe that they are on the continuum of primary treatment and should be part of the calculated time load on the PCP. Furthermore, some other important limitations exist.

First, an underlying assumption of use of the AADT is that a higher number of annual visits with a shorter average duration are equivalent to a lower number of annual visits with a longer average duration. If the first 2 or 3 min of each visit are used by the physician to greet the patient and look at the electronic notes of past visits, this may not be the case. In addition, these actions may require a minimum time allocated for each visit even when only one problem is raised. These issues are directly related to health care policy planning. Assuming there is a more efficient utilization of physician time with fewer yet longer visits, this aspect requires future examination, which could result in an organizational paradigm shift within the health care system.

Second, we excluded patients that died during the study year. We know that at the end of life the utilization of health care resources can be abundant [28–30], influencing the utilization of primary care visits as well. Therefore, to evaluate this special group, we will need another focused study.

Third, the analysis was not limited to one designated physician per patient, as it was designed to find the importance of the AADT required from primary care as a whole for the treatment of patients. This is an important aspect to be examined in future research, to investigate whether time spent with a patient’s personal primary care physician is more effective.

Another limitation of the current study is a possible information bias - some of those classified as “new immigrants” (as well as others) may live outside Israel. The fact that in recent years new immigrants to Israel usually keep their original residency increases the probability of such events.

Comparison with existing literature

As expected, chronic diseases were found to increase the number of annual visits with a PCP as well as the AADT. This coincides with previous research, which found patients with multiple chronic diseases having more outpatient visits per year, more adverse events, higher health care costs including the prescription of multiple medications, and having a lower health-related quality of life [31–33], This can be partially attributed to the fact that the average age and Charlson index score in our study were higher amongst patients with chronic diseases. In accordance with this finding, Østbye et al. found that chronic illnesses require more time then physicians have available for patient care [34].

In an overworked primary care system, facing growing numbers of elderly and chronically ill patients as well as mounting guidelines and tests, providing the required preventive, chronic and acute medicine and maintaining high quality of care is becoming an extremely difficult task [35].

To deal with these rising challenges on current models of primary health care, other forms of care such as shared medical appointments have been suggested [36]. This model of non-physician clinicians was also suggested by Yarnall et al., who proposed another solution in the form of many more shorter visits per year [37]. Additional recommendations include comprehensive primary care guidelines that integrate highly correlated diseases together, as well as patient education [34].