1. Background
Physician directories are among the most important databases in health systems. It shows the number and distribution of physicians in a distinct area. The ratio of different medical specialties could also be estimated using physician directories. Such databases could be used by governments and professional associations for medical human resource planning or national surveys (1-4). It could also be used by patients when they need to contact or have access to doctors (5).
There are two main physician directories in Iran: Islamic Republic of Iran Medical Council (IRIMC) (6) and the White Book (7). The first one is a free internet online database. The second is published in different ways as paper or electronic book, which has charge. In both databases, physicians are categorized by their specialties and sorted in each specialty by alphabetical order according to their last names.
The accuracy of physician databases has previously been questioned in other countries from different aspects. Some studies have only considered the accuracy of databases (8, 9), while others analyzed over- or under-estimation of existing physicians (10-12). In some other studies, the updating cycles of medical directories were also considered (13-15). Studying the quality, quantity and accuracy of information about the physician directories may lead to discovering the required changes for their improvement (16, 17).
2. Objectives
This study was designed for estimating the accuracy rate of physician directories in Tehran.
3. Materials and Methods
This was a cross-sectional study. Tehran, the capital of Iran and the concentration area of physicians, was selected as the study region. Afterwards, all medical specialists, fellowships and subspecialists of Tehran were extracted from IRIMC. For completing the physician's in-practice list, White Book was used as the additional source of information. This process was tended to prepare a complete list of physicians (specialists and subspecialists) of Tehran. Afterwards, cases with incomplete information (without phone numbers) or old phone numbers (less than eight digits) were eliminated from the list; thus, 5475 cases remained.
Telephone interviewing method was used for data gathering during 4-9 pm in all working days. The gathered data included the physician's name, sex, specialty, office phone number, and office address.
Contacts with no responses or busy numbers were reconnected during the week. Data gathering lasted six months, from July to December 2010. SPSS statistical software was used for data analysis.
4. Results
The information of 5475 physicians was extracted from the databases. Among the physicians, 24.1% (1320) were registered in IRIMC, 56.3% (3085) in White Book, and 19.5% (1070) in both.
Studying the gathered data showed that 14.0% (767) of numbers did not answer and the remaining 86.0% (4708) answered the phone calls; 26.8% (1467) of numbers were incorrect. The main reasons for existence of incorrect numbers were death, retirement or migration of physicians, and phone number changes. Of numbers, 59.2% were registered correctly; it means that the number was answered, it was correct, and the clinic was active (Table 1).
The accuracy rate was 59.2%. This rate was estimated 67.5% in IRIMC and 57.5% in White Book. Therefore, the accuracy rate was higher in IRIMC (P < 0.0001).
Table 1. The Frequencies of Answered (Incorrect and Correct) and no Answer Numbers in the Studied Databases
| Group | No. (%) | Percentage in Each Group |
|---|---|---|
| No answer | ||
| No response | 684 (12.5) | 89.2 |
| Busy phone number | 7 (0.1) | 0.9 |
| Disconnected phone line | 76 (1.4) | 9.9 |
| Total | 767 (14.0) | 100.0 |
| Incorrect numbers | ||
| Number change | 864 (15.8) | 58.9 |
| Address change | 290 (5.3) | 19.8 |
| Migration | 172 (3.1) | 11.7 |
| Death | 79 (1.5) | 5.4 |
| Retirement | 62 (1.1) | 4.2 |
| Total | 1467 (26.8) | 100.0 |
| Correct numbers | 3241 (59.2) | - |
| Total | 5475 (100.0) | - |
5. Discussion
Physician directories are important databases in health systems. They could be used in macro and micro level decision makings. In macro level, they are used for defining the number and distribution of physicians or handling national surveys. In micro level, they are useful for patients in selecting or finding their target specialists (1-5).
The study findings indicated that our databases did not adequately reflect all the physicians registered to practice in Tehran. The accuracy rate was estimated 60%. This is common in all databases with some differences. For example, the accuracy rate was estimated 90% in medical council of Jamaica (2667 registered and 2399 in practice physicians) (12). Nearly half of the cases registered in our databases in this study were incorrect. The main reasons of inaccuracy can be categorized in two groups: 1. There were physicians who were registered, but had migrated, been retired or dead. This rate was estimated 5.7% in our study. This probably caused over-estimation of the number of physicians. 2. There were physicians who were registered, but their phone number had been changed. This rate was estimated 21.1% according to the study findings. This failure may limit accessibility to physicians and lead to inaccurate estimation of the physicians' distributions. By the way, some cases did not answer, so we were not sure about their information accuracy; this rate was 14.0%. This may also limit the accessibility to physicians.
Lack of proper updating cycles may lead to these structural and functional defects. For example, American Medical Association (AMA) was not regularly updated with respect to retirement of physicians, causing overestimation of the physicians' numbers (8, 18). At first, the updating processes should be established; then, time cycles should be defined considering the characteristics of the system. Therefore, reviewing the workflow of each database is necessary. All physicians who want to practice through the country should register in IRIMC. Therefore, this database is well-designed for considering new entries. However, there is no defined process for eliminating dead, retired or migrated physicians from the list. This also applies for updating the phone numbers and addresses, since these data are only updated when physicians voluntarily refer to IRIMC for receiving some services.
The other studied database, White Book, is related to a nongovernmental publisher, established for commercial purposes. Hence, it may not consider all in-practice physicians and may underestimate their number. On the other hand, it was supposed to be revised yearly, but has not been updated so far; some structural and financial barriers have been the main causes of this negligence. This limitation decreased the database accuracy as well. Therefore, it has not been proposed for consideration in the structure of national medical directory.
The registration process of these databases, which was explained above, was inconsistent with the number of records in databases. It means that the entry process of IRIMC was more complete, but the frequency of registered cases in White Book was higher. More searches in IRIMC database showed that its access imitation (access to only 200 registered cases in each specialty according to alphabetical order) causes this matter. In fact, the information in this database is more than what is available. This limitation not only may cause under-estimation of in-practice physicians, but also will reduce the access to physicians' information at the end of the alphabetical list. Eliminating this limitation will increase the available cases in this database and turn IRIMC to the core national medical directory.
Both medical directories should try to define revision or updating phases for eliminating dead, retired or migrated physicians as well as updating the phone numbers and addresses periodically to increase the accuracy of information. However, it is beneficial to design a national network that connects different responsible organizations such as death registry, medical evaluation organizations, and medical sciences universities in Iran, to present their information about physicians' death, retirement or migration to medical directories, especially IRIMC. This can present a basic structure for national medical directories. This network will provide a continuous flow of information, which helps to update the physician directories.
