like with any correlation, there can be multiple reasons for the behavior you observe - first of all, did you try to vary the lag? I mean, if one takes a weekly averaged concentration of pollutants (A) vs number of admitted patients (B) and builds a linear correlation coefficient for A and B series, moving B forward in 1 week increments, one can detect a 2-3 week lag between the cause and the consequence. If you prove that the k_corr is maximal for, let's say, 3 week lag (and k_corr is large) and shows the same behavior each year then at least there will be a subject for the discussion.

If you talk about "concentration during monsoon" I guess you have data over a long period (maybe one year?) so a lag of few days and even weeks may not help.

I am not an expert on monsoon but admissions for pulmonary diseases may be driven by many factors including temperature, humidity and so on. Unless the pollutant concentration is high or very high, poor air quality may not be the first cause of hospital admission.

I would try to have data about two geographic areas with the same meteorology (and same socioeconomic conditions) but different pollutant concentration. If you observe the same admission rates then the impact of air quality may not be significant.

Parth sarathi, your question is very much useful and in public interest. Long time back we did some study and found good correlation between air contaminant data and hospital data. However, Central/ State , Pollution control boards, and National Environmental Engineering Research Institute, Nagpur and several  such agencies are doing such studies and they can be contacted. Based on contaminant data, near by hospital can also be checked. If not today, tomorrow such type of study will be very much required.

Sir, First of all you have to find out the reasons for pulmonary disorders. That largely depends on your area of work. Like in northern India, this kind of behaviour can be observed due to large impact of crop residue burning( which in your case may be due to pollution from nearby factory or any other mining activity). Plus meteorological factors play a great role in your case as generally all the pollutants are washed out from atmosphere in the rains or monsoon season, only met data can provide suitable information. I think you should do a ground survey first of local air pollution sources near that hospital , only than you will be able to correlate this uncommon behaviour. Hope this will be of little help:)

More generic way of arriving at dependency equations like correlation is based on MCDM, Multicriteria decision making. 

You may first resolve your objective into all those variables which affect your goal. i.e find all the critical parameters resulting in pulmonary diseases from health experts. Rank all of them in order of priority, as the case may be of immediate effect of parameter ( lag ).

correlate ( multi parameter )  these effects ( of each parameter ) along with air quality as one of them. This may give you some of the top ranking scenarios where in disease spread with and without inclusion of each of the combinations will result.

You may be get results ( ranking ) where in air quality is included, giving reasons of air quality induced disease prevalence.

All this efforts and exercise is because of non-linear, indirect dependency of causative factors towards the identified study.

Thank you Dr. Feofilov for giving such inputs. I would definitely try something like this. Would really be happy if you have any literature in mind.

Dear Dr. Finazzi, thank you very much for your suggestion and a comparative study really makes sense and maybe I would take up such studies in future also. However being a part of my Ph.D. work at present my study site and hospital records are limited.

So if at all you have any suggestions to explain such a variation would highly appreciate.

you are welcome :)

No, I didn't have anything specific in mind. The approach I was talking about is quite common and the libraries for correlation analysis exist in many programming languages.

If you want to dig further, you can search for principle component analysis books, but this might be an overkill for your problem. At the first glance, I would say that the correlation analysis with variable delay should give an answer.

Dear Dr. Saxena your encouragement means a lot for persuing such a study. I am trying my best to make some sense from these records and utilizing scientific data.

Dear Gurpreet thank you for your suggestions, I will surely consider the factors.

Dear Dr. Shekhar your suggestions are also nice. I will try to talk to the doctors and try to find out possible factors and work along your lines of suggestions. Thank you.

 Dear Parth, hope this might be of any help to you.

Parth Sarathi, we have carried out some study and tried to correlate with nearby hospital record but due to some or other reason that remained unfinished. However, Following paper may be useful :

VKSaxena. 1989, Air pollution studies within NGRI campus and its surrounding areas, Hyderabad. Indian Journal of Environmental Protection, volume, 9, No. 6, page 423-425.

I don't know if this applies where you are, but Dr. Weston A. Price (Nutrition and Physical Degeneration) found that heart attacks were more common during the winter months when Vitamin D and the other fat soluble vitamin levels were lower in milk and butter in this country (USA).  Do you have seasonal pasture growth?  If so, the growing season should provide the preventative nutrients in meat and dairy.